Study on the economic impact of Generative AI in the Music and Audiovisual industries - Jaunt

Study on the economic impact of Generative AI in the Music and Audiovisual industries

AI Summary

Bulleted

Text

Key Insights

🤖 Generative AI Overview: Explores the technology, creative industry applications, and ecosystem.
⚖️ Copyright Management: Discusses issues and legal actions regarding AI's use of copyrighted material.
📈 Market Growth Drivers: Identifies technological progress, creator economy growth, and consumer habits.
💰 Economic Impact by 2028: Projects market size, creator revenue loss, and AI service revenues in music and audiovisual.
🤝 Methodology of the Study: Explains the methodology used in the research.

#Musicindustry#Audiovisual#Markettrends#Aiincreativity#Creativity#Futureofmusic#Revenuestreams#Copyrightissues#Economicimpact#Generativeai

1
Study on the economic impact
of Generative AI in the Music and
Confidential PMP Strategy Audiovisual industries
Study on the economic impact
of Generative AI in the Music
and Audiovisual industries
Complete study
Current situation and 5-year perspective
November 2024
Confidential PMP Strategy

1/116

2
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
1 Introduction: Study context & objectives p.4
2 Generative AI overview p.10
What is Generative AI? p.11
Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? p.20
What are the issues at stake in terms of copyright management? p. 24
What are the main trends driving the growth of Generative AI in creation, today and by 2028? p.31
3 Economic impact in Music and Audiovisual creation p.39
What will be the economic impact of Generative AI in Music by 2028? – use cases and economic estimates p.51
What will be the economic impact of Generative AI in Audiovisual by 2028? – use cases and economic estimates p.72
Approach and methodology p.40
Content of the study

2/116

3
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Disclaimer
1
This Study was prepared by PMP Strategy, an independent strategy consulting firm mandated by
CISAC, to assess the economic impact of Generative AI on creation in the Music and Audiovisual
sectors.
2
The study provides PMP Strategy’s independent and objective view on the evolution and impact of
the use of Generative AI services on the two repertoires considered up to 2028. The historical figures
and forecast assumptions are based on market data, relevant benchmark and interviews with
industry experts: Collective Management organizations (CMOs), creators, tech players, producers,
publishers, DSPs, and institutional players representative of the two industries.
3
Inevitably, unanticipated events and circumstances may occur, and some of the assumptions used
to develop the forecasts may not be realized. Consequently, while we consider that the information
and opinion given in this Report are sound, PMP Strategy does not guarantee or warrant the
conclusions contained in the Report.
4
The Study is valid at the date of completion, which may fall prior to publication. The authors do not
take responsibility for any information or events after the Report’s delivery date which may affect its
contents.

3/116

4
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
1. Introduction
Context, objectives,
methodology

4/116

5
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
CISAC has commissioned PMP Strategy to assess the economic impact
of Generative AI on creation in the Music and Audiovisual sectors
Market
size 1
What will be the market size of Music and Audiovisual outputs generated
by AI in 5 years (2028)?
Revenue
loss 2 Potential cannibalisation of creator’s revenue streams due to the substitution
of human works by Gen AI outputs
What will be the associated loss of revenue for creators by 2028?
Market penetration and market value (on both B2C and B2B segments) of Gen AI outputs
Gen AI
services’
revenues
3 Revenues of Gen AI tools aimed at the general public and professionals, offering either
complete outputs generation and/or assistance in the creative process
What will be the revenues of Gen AI tools/services providers by 2028?
Introduction: Context, objectives, methodology

5/116

6
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The evaluation focuses on (1)the value of Gen AI outputs in the market, 2)
The evaluation focuses on(1) the associated impact on creators' revenues,
The evaluation focuses on (1) the revenues of the tools enabling outputs’ generation
Source: PMP Strategy analysis
1
2
3
CREATION DISTRIBUTION/BROADCASTING CONSUMPTION
Gen AI
tool/service
Gen AI service
user
Distributor /
Broadcaster
Subscription/pay
per act
Subscription /
pay per act /
advertising
Creators
Creators’
revenue loss 2
End-user
Collections
CMOs
Other revenue
streams
Introduction: Context, objectives, methodology
Legend Gen AI output flow Money flow
Gen AI tools
revenues’ 3
Gen AI Outputs’
market size 1

6/116

7
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Two-level impacts
analysis
Impact of market penetration by Gen AI outputs and impact on creators' revenues
(both in terms of revenue loss due to cannibalisation and of revenue opportunity)
Broad scope Broad scope of the study:
2 repertoires, international footprint
Involvement of
industry experts
Strong involvement of CMOs and representatives from the industries:
50 industry professionals interviewed or involved in workshops
Use cases & market
trends analysis
Detailed analysis of use cases and underlying factors / market trends
determining their evolution over the next 5 years
Transparent
methodology
Transparent methodology and assumptions built and validated
with CMOs and CISAC team
Quantitative and qualitative analysis, based on interviews and insights
from industry experts, existing studies / market data and workshops
Exhaustive
approach
The study aims at identifying the main applications of Gen AI
in these fields by 2028 and estimating their economic impact
Introduction: Context, objectives, methodology

7/116

8
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Introduction: Context, objectives, methodology
The study is based on experts' interviews, internal and external data analyses, and workshops
…interviews with industry
professionals from the Music,
Visual Arts and Audiovisual
sectors between July and
September 2024
+50
(Creators, Producers, Publishers,
Distributors, DSPs, CMOs, Tech & AI
companies, institutional players)
Expert interviews
Public and private players’
data sources
Data sources
• Market data
• Studies and panels on use cases
and trends in Generative AI
• Literature and main texts on
regulatory context and
copyrights issues
Workshop sessions with
CISAC members
and industry experts
8
Workshops

8/116

9
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
50 interviews were conducted with representative stakeholders from the 2 industries, across the
value chain
Introduction: Context, objectives, methodology
Music Audiovisual Other institutions

9/116

10
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
2. Generative AI
overview

10/116

11
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Generative AI
overview
What is Generative AI?
Who are Gen AI services’ providers in the field of creative
industries and how is the ecosystem structured?
What are the issues at stake in terms of copyright
management?
What are the main trends driving the growth of Generative AI in
creation, today and by 2028?
Generative
AI overview

11/116

12
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Artificial Intelligence (AI)
Machine learning
supervised or unsupervised
Deep learning
(1)Large language models, enabling to perform a variety of natural language processing
tasks (generate human-like text, classify text, answer questions, etc.)
Associated functionalities
What are the main characteristics & specificities
of each technology?
Ex. of applications
Ability given to machines to mimic human
intelligence and cognitive functions to
perform various tasks (problem solving,
learning)
n.a. (research field)
How do they translate
concretely?
Ability given to machines to learn from
structured datasets, without explicit
programming, to detect patterns and make
decisions/predictions
Customer
behaviour
prediction
Weather
forecasting
Ability given to machines to learn complex
patterns based on artificial neural
networks, from large and unstructured
datasets
Facial
recognition
Documents
reading
Ability given to machines, notably relying on
LLMs(1), to learn complex patterns to
generate new content: language, visual and
audio, etc.
Image/video
creation
Human-like
interaction
AI fields
What are the different AI technologies?
How do they interlock with each other?
1950’s
1970’s
1980’s
2000’s
2010’s
2020’s
Key events popularizing
AI technological
advances
1996
G. Kasparov, world’s chess
champion, was defeated
by Deep Blue, IBM’s
developed supercomputer
relying on machine
learning
2011
Watson, IBM’s developed
supercomputer relying on
deep learning, won 1st
place in US quiz show
Jeopardy! against historic
champions
2022
OpenAI publicly
launches ChatGPT, an
AI-powered chatbot
engaging in
conversational
dialogues and
providing responses
to user queries
Generative AI
Source: Specialized press, Stanford University, PMPS Analysis
Generative AI is the recent pinnacle of 50 years of progress in Artificial Intelligence
What is Generative AI?

12/116

13
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
3.1. Prompt
• Instruction of a prompt by the user to the model (in the form
of text, image, video or audio)
3.2. Prompt processing
• Processing of the prompt by the AI program, leveraging the
training phase
3.3. Output generation
• Generation of an output (new content) in the form of text,
visual or audio content
AI model production and use
Key steps for the development and use of a Generative AI model
Generative AI models leverage deep learning on large datasets to generate new content
(image, text, video, audio) upon the user’s instruction
Source: Specialized press, PMPS Analysis
1.1. Data collection:
• Selection of large datasets,
relevant to the type of output to
be generated by the program
• Necessary mass copying and
storage of data
1.2. Data preprocessing:
• Preparation of the raw data for
analysis (cleaning, normalising,
labelling, enhancing, etc.)
2.1. Model architecture building
• Selection and building of the model architecture (including
GANs, VAEs, transformer-based : see detail on next page)
2.2. Model training
• Training of the model, taught from the pre-processed
dataset
• Unsupervised learning
2.3. Model optimization
• Continuous/iterative performance evaluation
• Adjustments/refinement of parameters (to minimise
difference between the output and real data)
• Improvement of the model structure
AI model development
1 – Collection & preprocessing 2 – Training
1.1. Data (input) collection &
1.2. pre-processing
2.3 - Model
optimization
2.2 - Model
training
2.1 - Model architecture
building
3 – Content generation
3.1 - Prompt 3.2 – Prompt processing 3.3 - Output
generation
What is Generative AI?

13/116

14
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Most Gen AI programs today are based on 3 models - GANs, VAEs and transformer-based
models, with specific applications and benefits
Source: Specialized press, PMPS Analysis
• DeepArt.io: Transforms user-uploaded
photos into artwork in the style of famous
painters using GANs
• ThisPersonDoesNotExist.com: Generates
lifelike human faces that don't belong to
any real individuals
Involvement of two neural networks in GANs:
• The generator, which creates data (produces data
so convincing that the discriminator cannot
distinguish it from real data)
• The discriminator, which evaluates it (becoming
better at identifying fake data over time)
• Image creation
• Realistic photographs generation
• Art, and fashion designs
• Video game environments
• …
Generative
Adversarial
Networks
(GANs)
Description Main applications Examples of AI Services
The field is rapidly evolving, with new models being developed regularly – other models include autoregressive, diffusion models, RNNs, EBMs, and flow-based models
• Jukebox by OpenAI: Produces music in
various genres and styles by sampling
and processing audio in latent space
• AIVA (Artificial Intelligence Virtual Artist):
Composes original music scores suitable
for films, games, and other content
Two key phases in the VAEs’ generative model:
1. Encode input data into a latent space
2. Decode to generate new, similar data
Learning of complex data distributions and producing
new instances similar to the input data
• Image generation
• Synthetic datasets creation
• Drug discovery
• Music generation or other audio content
• …
Variational
Autoencoders
(VAEs)
Use of attention mechanisms to process sequences of
data (text or pixels), by focusing on different parts of
the data at different times
Generation of coherent and contextually relevant
content
• Natural language tasks (translation,
summarization, and text generation)
• Image-related tasks
• …
Transformerbased Models
• GPT-3 by OpenAI: An advanced language
model capable of understanding and
generating human-like text (answers to
questions and creates content)
• DALL-E by OpenAI: Generates imaginative
images and art from textual descriptions
What is Generative AI?

14/116

15
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Text-based prompt Audio + Lyrics
Generation process of Gen AI outputs
Generative AI engines can handle all data formats to generate increasingly diverse contents,
and perform a wide range of tasks
Source: Specialized press, Suno website, PMPS Analysis
3.1 - Prompt 3.3 - Output
generation
Example:
3.2 – Prompt processing
3.1 - Prompt 3.3 - Output
generation
3 – Content generation
3.2 – Prompt processing
3.1 - Prompt 3.3 - Output
generation
Question answering
Sentiment Analysis
Information
extraction
Image Captioning
Object Recognition
Instruction Following
Images-based prompts
Audio-based prompts
Text-based prompts
Mixed-media prompts
Text output
Image output
Audio output
Mixed-media output
Video output
>
Video-based prompts
Others
What is Generative AI?
Prompt format Task Output format

15/116

16
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Illustration of Midjourney technology performance evolution | 2022 – 2023
Recent cases have demonstrated the ability to generate content always closer to human
creations
Source: All Midjourney Versions (V1-V6) Compared: The Evolution of Midjourney - Aituts
Midjourney V1 Midjourney V2 Midjourney V3 Midjourney V4 Midjourney V5.1 Midjourney V6
Initial version with
raw results
Feb 2022 April 2022 July 2022 Nov 2022 Mai 2023
Introduction of
upscaling and
variations,
improved
coherence
Improved lighting,
reflections and
realism. Added
stylised and quality
parameters
Photorealistic
quality, ability to
generate complex
designs
V5.1 to V5.2:
Greater realism
and aesthetics
Text & Image-based prompt Image output
vintage photo, girl
smoking cigarette,
irina nordsol kuzmina,
a hazy memory, pixiv
--ar 2:3
Image Text prompt
Dec 2023
Improved image
quality and prompt
understanding
What is Generative AI?

16/116

17
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Extract of Qianqiu Shisong, China Media Group AI-generated animated series | 2022
Generative AI tools are thus increasingly questioning the very notion of creation
Source: Experts’ interviews, specialized press
• In February 2024, Chinese state broadcaster, China
Media Group (CMG), launched the country’s first
animated series created with a Generative AI tool,
Qianqiu Shisong, which features ancient stories
based on traditional Chinese poems and verses,
and aims to showcase the country’s traditional
culture and aesthetics.
• The series was produced using CMG’s internal
text-to-video model (Media GPT), trained on
traditional Chinese poetry and video and audio
material from China Media’s catalogue.
• The production studio indicated that artificial
intelligence was used at every step of the
development and production process, from
design to video generation and post-production.
What is Generative AI?

17/116

18
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
0.8
2.1
4.2
2.9
25.2
2019 2020 2021 2022 2023
22.5
0.7
0.6
1.4
2023
25,2
Private investment in Generative AI | Worldwide, 2029 – 2023, $Bn
While total investment in AI has recently slowed down, the Generative AI market is
skyrocketing, with an unprecedented surge in 2023 highly driven by the US
Source: Stanford University – AI index report 2024, PMPS Analysis
x9
Other
China
EU + UK
USA
Worldwide
• Investments in Gen AI have surged recently as
the technology demonstrates its potential to
transform industries and reshape the business
landscape
• A wide range of startups and Gen AI applications
are targeted by investments in sectors such as
technology, telecom, healthcare, financial services,
energy, consumer goods, media, culture, and
entertainment
• Generative AI is becoming a key driver of
innovation, with applications that enhance
operational processes and create new products
and services, impacting nearly every aspect of the
modern economy
Over the past few years, we've witnessed a significant
surge in investments in Generative AI by major tech
companies and private investors. This trend is driven
by the potential of Generative AI to revolutionize jobs in
various sectors.
Tech Company
>
What is Generative AI?

18/116

19
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Generative AI encompasses a fragmented and fast-developing ecosystem, with major
generalist players mostly related to GAFAM and multiple smaller solutions serving specific
purposes
Source: Specialized press, PMPS Analysis
The ecosystem has seen an exponential growth in the last year, and is polarized around a few mature and powerful big players, mainly related to GAFAM
(~1bn visitors/month on OpenAI.com), and a very scattered network of small and specialized newcomers
Generative AI services mapping (non-exhaustive)
Specialists
Generalists
Mature
GPT-3.5
Microsoft
(Bard) - Google
Microsoft
Newcomers
Amazon / Google
Elon Musk
Meta
Salesforce
Google
What is Generative AI?
19
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries

19/116

20
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
What is Generative AI?
Who are Gen AI services’ providers in the field of creative
industries and how is the ecosystem structured?
What are the issues at stake in terms of copyright
management?
What are the main trends driving the growth of Generative AI in
creation, today and by 2028?
Generative
AI overview

20/116

21
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Creative inspiration / Research and
planning Creation / Execution
In Music and Audiovisual, Gen AI services have emerged with use cases ranging from
assistance on specific tasks to fully automated complete outputs generation
Source: Experts’ interviews, specialized press
Post-production
(1) Mainly for general users (entertainment) | (2) Mainly for artists/professionals | (3) Not available yet
Audiovisual
Music
Full execution / composition(1) Support to creation(2)
iMyFone VoxBox
(3)
Gen AI ecosystem in creative industries

21/116

22
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The ecosystem in these fields is mainly made up of very recent, fast-growing newcomers(1)
Source: Specialized press, PMPS Analysis
Note: (1) As of July 2024
2016 2017 2018 2019 2020 2021 2022 2023 2024
Sora
Make-A-Video
by AudioCraft
by
• New ‘Suno for Mobile’ launch in July 2024
• 12 million total users (1)
• Approx.. 860k tracks generated/day (1)
Gen AI players (non-exhaustive)
Gen AI ecosystem in creative industries

22/116

23
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Models vary according to the use cases addressed and target audiences, and have not all yet
reached their full level of maturity
Source: Experts’ interviews, specialized press
Consumer Prosumer Professional
Main
Players
Complete prompt-to-output tools Assistance in the creative process
Destinated to mass public use Destinated to industry professionals
Note: (1) Experienced amateurs
• Mainly newcomers, tech startups
• Newcomers / tech startups
• New tools / functionalities of traditional players
Services
/ Key
features
• Mainly complete prompt-to-output generation
(images, video, music)
• Gen AI tools / functionalities
to automate specific tasks :,
mastering, editing, image
enhancement, …
• Prompt-to-output generation
• Gen AI tools / functionalities
to automate specific tasks
Economic
model/
offer
• Mainly free-use and freemium (subscription-based) models
• Revenues driven by ads, external investments and subscriptions
(freemium models)
• Mainly Licenses
(packages’ purchase)
models
• Free-use and freemium
(subscription-based) models
• Licenses models
StyleGAN
Consumer tools, also suitable for professional use Prosumer tools Professional tools/software
Gen AI ecosystem in creative industries

23/116

24
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
What is Generative AI?
Who are Gen AI services’ providers in the field of creative
industries and how is the ecosystem structured?
What are the issues at stake in terms of copyright
management?
What are the main trends driving the growth of Generative AI in
creation, today and by 2028?
Generative
AI overview

24/116

25
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The use of Gen AI in creative industries raises two main issues related to copyright
management
AI models’ inputs AI models’ outputs
• To what extent AI models have been trained by using
copyrighted works in their datasets?
• How can creators be remunerated for the use of their
works to train Gen AI models?
• What will be the implications of the increasing use of
synthetic data?
• Do Gen AI outputs infringe copyright on existing works?
• e.g., creation of works “in the style of”
• Who is/should be liable in case of copyright
infringement?
• What could be the ownership and “copyrightability”
of Gen AI content?
• What can/should be considered as Gen AI content?
• Should Gen AI outputs be protected? Who would
own the rights on Gen AI outputs?
Source: Experts’ interviews, specialized press, PMPS Analysis
Copyright issues & current regulatory framework

25/116

26
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
SUNO vs. JEN AI vs. Genario: a comparison of copyrights concerns along the Gen AI supply chain
There are widely varying degrees of transparency in the operation and construction of AI
models
Specialized press, PMPS Analysis
When asked about OpenAI’s training data sets, CTO Mira Murati responded: “We used publicly available data and licensed data”.
However, publicly available data doesn’t mean copyright-free data (ex: Youtube videos).
Source:
• No transparency on dataset
• No compensation / licensing of
the data used for training
• No publication or declarations on model functioning
• Outputs’ ownership to Suno for free subs.,
and to users for Pro/Premier subs
• No verification of the outputs
• Model trained on licensed
catalogues (Fairly Trained certif.)
• No licenses with creators’
representatives
• Publication of research paper on how the model works
• Outputs’ ownership to the user
• Audio recognition and copyright
identification of outputs
3 – Content generation
AI model Use
3.2 – prompt
processing
3.1 - Prompt 3.3 - Output
generation
1 – Data collection & preprocessing 2 – Model training
AI Model Development
1. Data (input) collection
& pre-processing
2.3 - Model
optimization
2.2 - Model
training
2.1 - Model architecture
building
• License with SACD (audiovisual
CMO) to remunerate authors for
the use of their works in the
models’ training
• Publication of research paper on how the model works
• Outputs’ exclusive ownership to the user
• No outputs used to train the model
Copyright issues & current regulatory framework

26/116

27
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The performance of Generative AI models is closely tied to their training datasets, ranging from
in-house data to AI-generated synthetic data, raising copyright concerns
Source: Experts’ Interviews, Specialized press, PMPS Analysis
Catalogues
& Players
(non-exhaustive)
Internal user generated content/data
In-house data
User-generated content
Used to tailor models to
specific user base
External data source
Public data/Open source
Used for the quantity
of resources available
Public datasets Open-source
repositories
Available via
Proprietary
Used for the quality of
the catalogue
Copyrighted works catalogues
AI-generated data source
Synthetic data source
AI-generated content
Used for the constrained-free offer :
logistics, privacy and copyright etc.
• As synthetic datasets become more widespread, transparency issues in the training process of Gen AI models are becoming increasingly urgent
• However, AI models will always need non-synthetic, human-made data for bias mitigation, renewed creativity and staying in touch with current trends
Specificities
& risks
• Highly relevant and personalized
datasets
• Enables to improve user
engagement and satisfaction
• Privacy concerns
• Data bias reflecting user base
• Broad coverage and cost effective
• Promotes transparency and
reproducibility
• Variable quality and reliability
• Potential for outdated or irrelevant
data
• High accuracy and reliability
• Domain-specific insights
• Expensive data
• Licensing restrictions
• Potential ethical concerns
• Cost and time efficient (generated
faster and more affordably)
• Non-copyrighted material
• Data availability
• Consistency and control
Copyright issues & current regulatory framework

27/116

28
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Industry players have started taking legal action against Gen AI services for the unauthorized
use of their catalogues in models training
Source: Specialized press, Music Business Worldwide, Experts’ interviews, PMPS Analysis
• In June 2024, Universal, Warner and Sony filed federal copyright infringement lawsuits
against AI music generator platforms Suno and Udio
• The Majors accuse the platforms of "mass infringement of copyrighted sound
recordings copied and exploited without permission.“
Major music groups vs. Suno &
Udio
VS.
1.
• The New York Times filed a lawsuit against OpenAI and Microsoft for copyright
infringement, claiming that millions of its articles were used without authorization
• These works were allegedly used to train AI technologies like ChatGPT, which now
compete with the newspaper as a source of reliable information
New York Times vs. ChatGPT
VS.
3.
• Universal Music, Concord and Abkco have filed a copyright infringement lawsuit against
the AI start-up Anthropic
• They accuse the AI company of using their artists' lyrics without permission to
generate near-identical copies through its AI model Claude
Majors & music groups vs.
Anthropic
2.
VS.
• A trio of artists (Sarah Andersen, Kelly McKernan and Karla Ortiz) filed a lawsuit against AI
companies Stability AI and Midjourney for copyright infringement
• Several artists have joined the legal action, including other AI services in the prosecution:
DeviantArt and Runway AI
4. Artists vs. Stability & Midjourney
VS.
Other
examples
Music
• In addition to legal action taken against Gen AI services, CMOs have started to establish opt-out mechanisms to prevent the training of Generative AI models using
copyrighted works of their members (e.g., Sacem for Music)
• However, these mechanisms only apply to future training of AI models and are made possible by laws mandating transparency in the training processes of AI models
Copyright issues & current regulatory framework

28/116

29
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
United States Europe Rest of the world
Overall, the regulatory framework is still in progress and remains heterogeneous across regions
Source: Expert interviews, specialized press
Generative AI Copyright Disclosure Act
(April 2024):
A proposed law requiring AI companies to submit a
list of all copyrighted works used for training their AI
models.
US Copyright Office Guidelines (March 16, 2023):
Clarified the necessity of human contribution to
qualify for copyright protection, stressing that tools
like AI can be part of the creative process, but human
control over the expression is essential.
Directive on copyright and neighbouring rights
in the digital market (April 2019):
Directive allowing text and data mining (TDM)
necessary for AI training under certain conditions:
• Article 3: Allows data mining for scientific
purposes without special conditions
• Article 4: Allows data mining for all purposes,
including commercial, provided access to the
data was lawful and rightsholders did not opt out
AI Act (April 16, 2024):
Introduces several obligations for AI systems:
• Ensuring respect of copyright, including for opensource foundational models
• Publishing detailed summaries of works used for
AI training
• Identifying AI-generated content as such
• Extraterritorial application, effective from August
1, 2024, with phased implementation until 2027
Council of Europe Framework Convention
on AI (May 17, 2024):
Focuses on respecting human rights in AI
development, emphasizing transparency for
enforcing intellectual property rights.
Guidance for Gen AI in education and research
(UNESCO, Sept. 2023):
Calls for immediate actions and long-term policies
to regulate the use of Gen AI in education & research,
including text, image, video and music generation.
Copyright Act – Art. 30-4 (Japan, May 2018):
Copyrighted works can be used in the training of AI
models without requiring a license. Rightsholders do
not have the option to opt out, and there is no
obligation for transparency
Gen AI governance framework
(Singapore, March 2024):
Advises policymakers to clarify the application of
existing personal data laws to Generative AI. Aims to
foster trusted Generative AI development.
AI policy on regulations & ethics (Israel , 2023):
Focuses on responsible AI innovation. Emphasizes on
“soft regulation” with sector-specific guidelines. Aims
to respect the rule of law, fundamental rights and
public interests.

Non-exhaustive list
Copyright issues & current regulatory framework
Legend
Mainly focused on input issues
Mainly focused on output issues

29/116

30
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Tools are also being developed to support industry players by helping them identify
copyrighted works used as inputs, and detect Generative AI outputs
Source: Expert interviews, Specialized press
Example of tools
Deezer is currently developing tools to:
• Identify whether a Gen AI model has been trained on specific tracks
• Detect music tracks generated by the world's largest
music-generating LLMs
Detection of copyrighted works used as inputs
Objective: Analyse Gen AI models to detect if copyrighted content has
been used in their training
Method: Compel models to provide specific copyrighted works as outputs,
with prompts designed to induce hallucinations, thus proving they have
been used in the training process
Challenges: Difficult to scale/industrialize the process
Detection of outputs generated by AI tools
Objective: Scan specific works or entire catalogues to identify whether they
have been generated by AI
Method: Identify Gen AI models biases and patterns (usually specific to
each model) and scan images/music/video catalogues to identify
whether they have been AI-generated
Challenges: Detection tools need to be regularly trained on popular Gen AI
models to ensure they remain performant
Spawning AI is developing solutions to help
identify whether a visual work has been used as
Gen AI tools inputs (Have I been Trained?), help
block AI web scrapping and enforce opt outs
Ircam Amplify has developed a tool
(AI-Generated Detector) allowing to
identify and tag Gen AI musical
outputs
Copyright issues & current regulatory framework
AI models’ inputs AI models’ outputs

30/116

31
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
What is Generative AI?
Who are Gen AI services’ providers in the field of creative
industries and how is the ecosystem structured?
What are the issues at stake in terms of copyright
management?
What are the main trends driving the growth of Generative AI in
creation, today and by 2028?
Generative
AI overview

31/116

32
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The adoption of Generative AI tools and outputs in creative industries will be determined by the
strategies and behaviour of players across the value chain
Source: interviews with industry professionals, Specialized press, PMP Strategy analyses
Main drivers Impact on Gen AI adoption Positioning in the
value chain
Market drivers and underlying trends
Technological
progress
• Rapid technical evolution of Generative AI models is expected to continue in the coming years, enhancing their
capabilities beyond simple text or image generation to more complex, multi-modal outputs
• Higher quality, more diverse, and personalized outputs, opening new opportunities across multiple industries
Gen AI
providers/Tech
companies
Growth
of the creator
economy
• Continued growth of user-generated content on social media, fostering the adoption of Gen AI tools to support
content creators
• Further reduction of the barriers to entry for creation in all creative industries driven by Gen AI tools
Creators
Evolution of
consumer habits
• Growing demand for interactive, on-demand, and contextually relevant content reshaping consumption
• Increasing trend toward passive content consumption, where digital platforms curate and recommend content to
users rather than users actively selecting it themselves, with Gen AI likely to play a pivotal role in driving this shift
Consumers/
End-users
Strategy and
positioning of
traditional players
• Shift in traditional players' strategies and positioning to adopt Generative AI for competitiveness, in all industries
• Integration of Generative AI by players across all segments of the creative industries' value chain: to introduce new
offers, optimise content production and distribution, and renew business models
Distributors/
Broadcasters &
B2B players
Regulatory
environment and
ethical issues
• Evolving regulatory frameworks addressing intellectual property, data privacy, ethics, and cultural diversity issues
as Gen AI becomes more widespread, potentially impacting its growth
• Increased awareness of end users regarding ethical issues related to copyright, fair pricing, and the proper
remuneration of authors
Legal
bodies/CMOs

32/116

33
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Technological
progress
Test scores of AI systems on various capabilities relative to human performance | 1998 - 2022
Recent technological progress have enabled AI and Gen AI models to outperform human
performance in all basic capabilities, laying the foundation for continued progress in the
coming years
Source: Kiela et al. (2023) & Our World in Data, PMP Strategy analysis
Within each domain, the initial performance of the AI is set to –100. Human performance is used as a baseline, set to zero.
• AI performances are above human
for every basic (non-complex)
capabilities analysed
• The later a capability started to be
implemented, the faster it reached
human-level performance:
– Speech recognition: 19 years
– Handwriting recognition:
17 years
– Image recognition: 7 years
– Reading comprehension: 1 year
– Language recognition:
less than 1 year
• Technological advancements have
paved the way for Generative AI to
revolutionize various sectors and
domains, including the creative
industries
The growth of Generative AI is expected to accelerate even further in the coming years, with widespread adoption and advanced
technical capabilities anticipated by 2028, driven by substantial investments in Gen AI models and their associated providers
Market drivers and underlying trends
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022
Human performance
Reading
comprehension
Language
recognition
Image
recognition
Speech
recognition
Handwriting
recognition

33/116

34
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Social media penetration worldwide and creator economy
The significant increase in user-generated content on social media will likely drive the high
adoption of Gen AI
Source: Specialized press, Experts’ interviews, PMPS Analysis
0%
5%
13%
28% 31%
37%
42% 45% 48%
53%
58% 59% 62%
2000 2005 2010 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
3.05
2.49
2.00
2.00
1.56
• The widespread adoption of social
media has driven the growth of the
creator economy across all
disciplines
• The creator economy has been
boosted mainly by influencer
marketing, content monetization,
e-commerce and social selling
• Approximately 15% of the 4.2Bn
users of social media (both
individuals and businesses/brands)
worldwide are considered part of
the creator economy
• These creators, and particularly
enterprises, are more inclined to
use Gen AI tools to produce more
personalized content for their users
at scale
– e.g., Nike and Coca Cola using
Gen AI algorithms to create
personalized and engaging
advertising campaigns
Social media penetration rate | %, Worldwide, 2000-2024 Top 5 social media in number of
users | in Bn users
Number of players in the creator economy in social media & split by type of players
4.2Bn
Social media users
Creators
Use their influence to aggregate
and monetize their audience
Creator economy = c.15%
of the social media users
Of which 2/3 are brands/enterprises
(vs. individuals)
Passion economy users
Engage in any activity to monetize skills
on digital platforms/social media
Market drivers and underlying trends
500m
200m
Creator
economy’s growth

34/116

35
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approximate average number of tracks uploaded to DSPs each day | thousands, per year, 2018 - 2024
New tools, formats, and distribution channels have significantly lowered barriers
to entry in music creation in the last decade, a trend which will be further fostered
by Gen AI tools
Source: Music Business Worldwide, PMPS Analysis
• Music streaming platforms have
favoured the creator’s economy in
the music field, with:
– Lower barriers to entry: easier
distribution for independent
artists without the need of
traditional record label
– Increased visibility and reach,
through algorithms and curated
playlists
– Monetization opportunities,
through the creation of a new
revenue stream with the
streaming royalties
– Access to data & analytics,
community & networking,
creative freedom…
20
40
50
60
100
125
143
2018 2019 2020 2021 2022 2023 2024
x3
x2.5
Music streaming providers (non-exhaustive)
The streaming platforms and other digital players (e.g., music distributors like CD Baby), combined with the
advancements in Generative AI tools, are empowering artists to produce, distribute, and monetize their music
more easily than ever before
Market drivers and underlying trends Creator
economy’s growth

35/116

36
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Passive music consumption (mood playlists) in streaming platforms | Focus on Spotify
Growing demand for interactive, on-demand, contextually relevant content, and the increasing
trend toward passive consumption is driving the adoption of Gen AI by streaming platforms
(1/2)
Source: Music Business Worldwide, PMPS Analysis
• Music streaming platforms
increasingly offer curated playlists
tailored to specific moods and
activities, enhancing user
engagement
• These mood playlists provide a
seamless listening experience that
requires minimal intervention from
the user
• End-consumers on streaming
platforms are increasingly
gravitating towards passive music
listening, driven by convenience,
personalization, and discovery
• In the top 100 Spotify playlists in
terms of subscribers, 41% are
considered as functional/mood
playlists (e.g., Morning Coffee),
favouring passive listening
Rank Playlists’ name # of subs
1 Today's Top Hits 34M
2 Top 50 - Global 17M
3 RapCaviar 15M
4 Viva Latino 14M
5 Rock Classics 11M
6 Baila Reggaeton 10M
7 All Out 2000s 10M
8
Songs to Sing in the
Car 10M
9 All Out 80s 10M
10 Beast Mode 10M
Rank Playlists’ name # of subs
89 This Is Michael
Jackson 3M
90 This Is One Direction 3M
91 Deep House Relax 3M
92 Hype 3M
93 Hot Hits Deutschland 3M
94 Warm Fuzzy Feeling 3M
95 Coffee Table Jazz 2M
96 Power Hour 2M
97 Intense Studying 2M
98 Chill Vibes 2M
Legend Mood playlists/passive listening
41%
59%
Commercial music
Mood music
Market drivers and underlying trends Evolution in
consumer habits

36/116

37
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Video streaming recommendation | Example of Netflix
Growing demand for interactive, on-demand, contextually relevant content, and the increasing
trend toward passive consumption is driving the adoption of Gen AI by streaming platforms
(2/2)
Source: Netflix reports, Specialized press, PMPS Analysis
• Consumers' appetite for recommended and personalised content is a key driver for the use of Generative AI,
enabling even more advanced playlists and content with unlimited tailored content, down to the individual level
Diagram on Netflix recommendation system : Collaborative filtering vs. content based
Collaborative filtering Content based recommendation
Similar users
1
2
Liked by both
Liked by her Recommended
to him
Similar content
1
2
Liked by her
Recommended
to her
• Netflix recommendation’s algorithm
includes :
– Data collection: viewing history,
user interactions, and
demographic data
– Collaborative filtering: user-user
and item-item
– Machine learning models :
identification of pattern in the user
preferences
– Content analysis : metadata
analyses, natural language
processing
– Real-time personalization
• This directly influences user
consumption behaviour: increased
engagement and enhanced user
experience
Market drivers and underlying trends Evolution in
consumer habits
of Netflix viewership is driven by its recommendation engine 75%

37/116

38
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Examples of current and projected Gen AI use cases across the main segments of the creative industries' value chain
Stakeholders across the entire value chain of the creative industries are increasingly adopting
Gen AI to optimize content production and distribution and to renew their value proposition
Source: Specialized press, Experts’ interviews, PMPS Analysis
• Music: Audiovisual production companies using Gen AI to
produce background scores in audiovisual content, lowering
production costs, mainly in lower budget works
• Audiovisual: Brands using Gen AI video outputs on social
media, enhancing end-consumer experience with more
tailored and personalized content
B2B distributors
• Music: Streaming platforms integrating Gen AI outputs in
mood playlists, to create more tailored content with no
copyright
• Audiovisual: VOD platforms using Gen AI videos to create
trailers, cutting production costs
B2C distributors
• Music: Press agencies using Gen AI to create background
scores, lowering production costs
• Audiovisual: Advertising agencies using Gen AI to create
personalized video ads, reducing costs and time
Artworks/Content aggregators Commissioners of artworks/content
• Music: Music libraries using Gen AI to generate large number
of new tracks, increasing catalog options and reducing
prices
• Audiovisual: Stock photo/video agencies using Gen AI to
create short videos content, cutting costs and speeding up
availability
Traditional
Market drivers and underlying trends players’ strategies

38/116

39
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
3. Economic impact
in Music and
Audiovisual
creation

39/116

40
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approach and Methodology
What will be the economic impact of Generative AI in Music by
2028?
What will be the economic impact of Generative AI in
Audiovisual by 2028?
Gen AI
Economic
impact

40/116

41
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Key figure
1
Key figure
2
Key figure
3
Market
size 1
What will be the market size of Music and Audiovisual outputs
generated by AI in 5 years (2028)?
Revenue
loss 2 Potential cannibalisation of creator’s revenue streams due to the substitution
of human works by Gen AI outputs
What will be the associated loss of revenue for creators by 2028?
Market penetration and market value (on both B2C and B2B segments) of Gen AI outputs
Gen AI
services’
revenues
3 Revenues of Gen AI tools aimed at the general public and professionals, offering either
complete outputs generation and/or assistance in the creative process
What will be the revenues of Gen AI tools/services providers by 2028?
Approach and methodology
CISAC has commissioned PMP Strategy to assess the economic impact
of Generative AI on creation in the Music and Audiovisual sectors

41/116

42
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries Source: PMP Strategy analysis
CREATION DISTRIBUTION/BROADCASTING CONSUMPTION
Gen AI
tool/service
Gen AI service
user
Distributor /
Broadcaster
Subscription/pay
per act
Subscription /
pay per act /
advertising
Creators
Creators’
revenue loss 2
End-user
Collections
CMOs
Other revenue
streams
Legend Gen AI output flow Money flow
Gen AI tools
revenues’ 3
Gen AI Outputs’
market size 1
Approach and methodology
The evaluation focuses on (1)the value of Gen AI outputs in the market, 2)
The evaluation focuses on(1) the associated impact on creators' revenues,
The evaluation focuses on (1) the revenues of the tools enabling outputs’ generation
1
2
3

42/116

43
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The methodology relied on qualitative and quantitative analyses,
fuelled by interviews with industry players and workshops with CMOs
Source: PMP Strategy analysis
Use cases identification
and prioritisation
Qualitative approach
• Qualitative analysis used to feed the quantitative part
(market hypotheses and impact estimates)
• Identification of market segments where generative AI
has a significant impact
• Translation of qualitative assessments into quantitative estimates
• Identification of the most significant generative AI use cases
for creation in both fields
• Prioritisation of these use cases based on their potential adoption level
to determine the ones with the most significant economic impact
Economic impact estimation
Quantitative approach
Approach and methodology

43/116

44
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music
Generative AI use cases examples in Music and Audiovisual (non-exhaustive)
Generative AI has numerous applications in the filed of creative industries and can intervene at
all stages of the creative process, from ideation to post-production
Source: Experts’ interviews, specialized press, PMPS Analysis
Creative inspiration Research & Planning Creation/execution Post-production Distribution & diffusion
• Melody generation
• Lyrics creation
• Trend analysis
• Sample discovery
• Collaborative
composition
• Arrangement
optimization
• Prompt-to-music
generation through AI
• Automated mixing
and mastering
• Pitch correction
• Voice cloning,
synthesis
• Playlist
recommendation/
optimization
• Targeted marketing
• Screenwriting
exploration
• Storyboard generation
• Scene visualization
• Script analysis
• Animation generation
• Special effects creation
• Video editing
• Colour correction
• Translating/Adapting
• Content
recommendation
• Audience analysis
Scope of the study
Audiovisual
Approach and methodology

44/116

45
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
More generally, use cases fall into 2 categories: fully automated prompt-to-output applications
or assistance in the creative process
Source: Experts’ interviews, specialized press, PMPS Analysis
Fully Gen AI outputs
• Fully automated generation of musical or audiovisual
outputs via the use of Gen AI services
• No human input beyond the prompt (or marginal
intervention)
• Creation of musical, visual or audiovisual works with
the assistance of Gen AI tools, enhancing human work
• Significant human involvement in the creative process
(“augmented artist”)
AI-assisted work creation
> • Prompt-to-video tools such as Sora (OpenAI) or InVideo
• Actor rejuvenation, video restoration & colouring, sound,
digitalisation, etc. (Respeecher)
• Prompt-to-script tools (Genario)
• Automated dubbing-subtitling (Veed)
> • Prompt-to-song tools such as Suno or Udio • Pitch correction, editing, mastering, etc. (AudioShake,
iZotope)
• In the market size calculation, only fully Gen AI outputs are considered, as their distribution will affect the market by replacing human-created works.
• For the creators’ loss calculation, (i) In the music sector, fully AI-generated outputs will cannibalize creators' revenues in specific market segments; (ii) whereas in the
audiovisual sector, complete AI outputs and reduced production budgets due to Gen AI tools (e.g., screenwriting, translation) will lead to revenue losses
• There is a grey area where semi-automated works may still be considered as human creations. The study does not aim at estimating the Gen AI contribution in human works
Grey area
Approach and methodology

45/116

46
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
▪ What will be the Music and
Audiovisual market segments
impacted by Gen AI use cases in the
next 5 years?
▪ What will be the penetration rate and
market value of Gen AI outputs in 5
years?
▪ Can we expect an "AI boost"
(additional growth) due to Gen AI?
▪ What will be the share of existing
players’ (distributors) revenues driven
by Gen AI outputs?
Gen AI impact estimation methodology: Market size of Gen AI outputs in 2028
What will be the
market size of
Music and
Audiovisual
outputs generated
by AI in 5 years
(2028)?
i.e. market penetration
and market value of
Gen AI outputs
1 Market size
Key questions to be answered Calculation methodology
Source: PMP Strategy analysis
• Segmentation of the Music and Audiovisual distribution markets (both
B2C and B2B – including new Gen AI based services and current distributors)
• Estimate of 2023 market size for all the distribution segments likely to be
impacted by Gen AI outputs
• Forecast to 2028 based on historical growth and market trends
• Estimate of Gen AI outputs’ penetration rate for each segment in 2028
based on the prioritised use cases
➢1 Market size of Fully Gen AI outputs by 2028
Approach and methodology

46/116

47
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
▪ What would be the evolution of
creators’ revenues in the next 5 years
without Gen AI? - based on current
remuneration rules and historical trends
▪ What will be the share of this revenue at
risk due to the cannibalisation or
substitution of human-made works by
Gen AI outputs?
Gen AI impact estimation methodology: Creators revenue loss due to Gen AI cannibalisation
Revenue
loss 2
What will be
the associated
loss of revenue
for creators by 2028?
i.e. risk of cannibalisation
of creators’ traditional
revenue streams
Key questions to be answered Calculation methodology
Source: PMP Strategy analysis
CMO-collected revenues (for both repertoires)
• Breakdown of CISAC collections in segments and sub-segments
• Estimate of 2023 revenues for each sub-segment and forecast to 2028
• Estimate of cannibalisation rates due to Gen AI outputs by sub-segment
Other revenues (only for Audiovisual)
• Share of the production budget / dubbing-subtitling market going to
audiovisual creators/authors
• Estimate of cannibalisation rates due to Gen AI outputs by type of author
Approach and methodology
2➢ Potential revenue loss for creators by ‘28 compared to a no Gen AI scenario

47/116

48
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Gen AI impact estimation methodology: Creators revenue loss due to Gen AI cannibalisation
Revenue
loss 2
What will be the
associated loss of
revenue for creators
by 2028?
i.e. risk of cannibalisation
of creators’ traditional
revenue streams
Source: PMP Strategy analysis
Creators’ revenue streams considered for Music and Audiovisual repertoires
• The perimeter of rights managed by CMOs is very
heterogeneous between regions/geographies
• Only in a few countries do CMOs account for a large share
of creators’ revenues
• For the Audiovisual field, the scope has been extended to
capture a better proportion of creators/authors’ revenues
Audiovisual
creators’
revenue split
Music creators’
revenue split
• The perimeter of rights managed by CMOs is very
homogenous between regions/geographies
• CMOs collection account for a significant share
of creators’ revenues
• For the Music field, the scope for the revenue loss
calculation is hence the CMO-collected revenues
Approach and methodology
CMO-collected rights
Impact of Gen AI on CMO-collected rights
Rights from upfront payments and other revenues
Gen AI impact on upfront payments and other revenues
Legend

48/116

49
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
▪ What will be the evolution of the
Gen AI ecosystem in the Music
and Audiovisual fields by 2028?
▪ What will be the market penetration
of AI-assisted music and
audiovisual/video creation tools
among professionals by 2028?
▪ How many fully Gen AI promptto-outputs tools will exist by 2028, and
what will be their user base and pricing?
▪ What will be the overall revenue
generated by both AI assistance and
fully prompt-to-output tools by 2028?
Gen AI impact estimation methodology: Revenues of Gen AI services
Gen AI
providers’
revenues
3
What will be the
revenues of Gen
AI tools/services
providers by
2028?
Key questions to be answered Calculation methodology
Estimated revenue of Gen AI tools and services by 2028
Source: PMP Strategy analysis
Approach and methodology
AI-assisted creation tools
• Estimate of the professional Music and Audiovisual software markets
(editing, post-production…) in 2023, forecast to 2028
• Gen AI penetration rate on this segment
Full prompt-to-outputs tools
• Number of services providing fully automated prompt-to-music tools
• Forecast of the average number of users and average revenue per user
to 2028
3

49/116

50
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Gen AI impact estimation methodology: Revenues of by Gen AI services
Gen AI
providers’
revenues
3
What will be the
revenues of Gen
AI tools/services
providers by
2028?
Gen AI tools and service providers have been split in 2 categories
Source: PMP Strategy analysis
• Prompt-to-video complete outputs generator
• Prompt-to dubbing Gen AI providers
• Prompt-to-scripts Gen AI providers
• Gen AI tools for video ideation, mastering,
editing, post-production…
• Includes Gen AI tools providing prompt-tovideos as one of their services, but mainly for
professionals
Audiovisual
Full prompt-to- outputs tools AI-assisted creation tools
• Prompt-to-songs generator
• Gen AI tools for music ideation,
mastering, editing, post-production…
• Includes Gen AI tools providing
prompt-to-music as one of their
services, but mainly for professionals
Music
Approach and methodology

50/116

51
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approach and Methodology
What will be the economic impact of Generative AI in the Music
field by 2028?
Main Applications
What will be the economic impact of Generative AI in
Audiovisual by 2028?
2028 forecast
Economic
impact
in Music
creation

51/116

52
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Identification of Gen AI main applications in the music creation process
Source: Experts’ interviews, specialized press, PMP Strategy analysis
Creative inspiration / Research and planning Creation / Execution Post-production
Main use cases
AI-assisted or fully AI-generated
creation/composition AI-assisted post-production AI-assisted music conceptualization& idea generation
AI-assisted idea generation: Complete / partial
musical work creation (samples, moods, rhythms,
etc.) used as inspiration for human created work
AI-assisted Voice Cloning to create original musical
works
Voice synthesis/cloning
Out of scope
(linked to performers’ rights mostly)
AI-assisted Pitch correction,
Editing, Mastering
Automated complete musical output generation for core
content in audiovisual works (e.g., video games main
theme)
Automated complete musical output generation for
background content (in AV works, in-store sonorization,
social media…)
Automated complete musical output generation for
commercial distribution on music streaming platforms 5
6
7
Automated complete musical output generation and
consumption by end users for entertainment purpose 4
1
AI-assisted Melody creation & Lyrics generation
AI-assisted Orchestration, Instrumentation and
structure
2
3
9
AI-assisted Enhancement/enrichment of existing/humancreated music (e.g., verses addition) 8
Gen AI economic impact - Use cases Music

52/116

53
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Prioritisation of use cases based on expected impact on creators’ revenues and adoption probability – Matrix Analysis
Prioritisation of Gen AI use cases in the Music field
Source: PMP Strategy Analysis
Automated complete musical output generation for
commercial distribution (e.g. “mood playlists”) 5
Automated complete musical output generation for background
content (in AV works, in-store sonorization, social media…) 7
AI-assisted Idea Generation: Complete/partial musical
work creation (samples, moods, rhythms, etc.) 1
AI-assisted Enhancement/enrichment
of existing/human-created music (e.g., verses addition) 8
AI-assisted Melody creation & Lyrics generation 2
AI-assisted Orchestration, Instrumentation and
structure 3
AI-assisted Pitch correction, Editing,
Mastering 9
Very Low
Very High
Adoption probability
Very Low
Impact on creators' revenues Very High
Automated complete musical output generation for core
content in audiovisual works (e.g., video games main theme) 6
Automated complete musical output generation
for sole entertainment purpose (non-commercial use) 4
Main use cases
High adoption
Low impact
Mid adoption
Mid to low impact
Mid Impact
Low adoption
Legend: use cases impact on creators’ revenues
- Marginal - Moderate - High to very high
Gen AI economic impact - Use cases
53
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music

53/116

54
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Transformation of music streaming: end-users become music
curators
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Technology & Output Quality: The tools are
user-friendly and intuitive, but the quality of the
outputs is still limited compared to traditional
commercial music
• Usage/Adoption: Usage remains primarily
occasional and for entertainment purposes
• Techno & quality of outputs: Improvement of the
technology leading to increasingly higher-quality
outputs
• Usage/adoption: Widespread adoption and a
shift from occasional, ad hoc use to regular use,
similar to traditional streaming platforms
With the perfecting and democratization of text-to-song and voice cloning tools, endconsumers are moving from simple users to music curators, thus questioning the very
notion creator.
Tech company
Current
application
• Tools like Suno Audio, designed for the general public,
allow users to generate and listen to music tracks
created from a simple text prompt.
• For now, these tools are mainly used on an ad hoc basis
for entertainment purposes among friends, colleagues
etc.
2028
potential
application
Two possible scenarios :
1. Tools like Suno and Udio evolve to become new players
in the music streaming industry
2. Existing music streaming platforms integrate these
new AI-powered content generation features
themselves
Example of service providers
Example
AI-powered platform allowing end-users to both produce and listen
to AI-generated music (Suno app, see next page)
2028 est. level of
adoption/maturity
No direct impact but dilution of
human-created tracks in the overall
revenues of streaming
2 Revenue loss
Direct revenues for Gen AI providers,
either directly providing services, or
internalized in DSPs
Gen AI
providers’
revenues
3
Market boost : monetization of these
new features (for DSPs) or services
(for AI tech companies)
1
Market size /
Gen AI
penetration
Low Mid High
Legend
Automated complete musical output generation
for sole entertainment purpose (non-commercial use) 4
2028 main economic impacts identified 2028 impact on
creators’ revenues
54
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music Gen AI economic impact - Use cases

54/116

55
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Suno launched its ‘Suno for mobile’ app in July 2024, offering enhanced functionalities
Example - Suno’s latest app launch transforms passive music
listening into an interactive experience
Source: Suno.com website
The Suno mobile app allows users to
create and share music in new and
innovative ways.
The key functionalities of the app
include:
1. Music Creation
• Text-to-song : users can
generate songs by inputting
lyrics or descriptions
• Audio recordings : The app
allows the user to record an
audio and use it for the
song
2. Music Streaming
• Music curation : The app
provides tools to curate and
collect music that the user
enjoys from other creators
‘Suno for mobile’ promotion – Suno.com
I suspect that the strategy of both Suno and Udio is to become new streaming platforms. […] where users can engage [with the content],
create their own versions of it, republish it and become curators and creators themselves.
– Music and AI expert
Automated complete musical output generation
for sole entertainment purpose (non-commercial use) 4
Gen AI economic impact - Use cases Music

55/116

56
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
2028 est. level of
adoption/maturity
Penetration of AI-generated music on music streaming platforms
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Current tools can
already produce good quality music for such
purposes
• Usage/adoption: Usage is still limited but the
adoption remains difficult to quantify
• Techno & quality of outputs: Improved
technology with increasingly higher-quality
outputs
• Usage/adoption: high adoption potential for
functional music and “passive” listening, for both
individual and corporate customers
Generative AI represents a potential opportunity for DSPs to generate royalty-free
tracks and integrate them into their playlists. This approach could significantly
boost their margins by drastically reducing copyright costs.
Tech company
Current
application
• Generative AI tools are already being used to create full
music tracks for mainstream distribution on streaming
platforms, often included in functional playlists.
• AI-generated music could represent a significant
portion of mainstream music, particularly in functional
music and passive listening through suggested playlists
(mood/contextual playlists)
• DSPs might even use AI themselves to generate tracks,
create and curate playlists based on user preferences
and moods
Automated complete musical output
generation for commercial distribution
(e.g. “mood playlists”)
5
Example
AI-generated tracks produced to feed a DSP’s contextual “Morning
Motivation” or “Casual Run” playlist
Example of service providers
2028
potential
application
Low Mid High
Legend
2028 Main economic impacts identified
High potential cannibalisation of
music creators’ streaming revenues
2 Revenue loss
Revenues driven by subscription
fees from prompters, or by the
internalization into DSPs
Gen AI
providers’
revenues
3
Moderate penetration rate in volume
and value on streaming platforms,
particularly in mood playlists
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact - Use cases
56
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music

56/116

57
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Spotify’s catalogue now includes AI-generated music created and uploaded by third parties
Example – 100% AI-generated music is already streamed on DSPs
Source: Spotify, Boomy, Expert interviews, Specialized press, Music Business Worldwide
• AI-generated tracks are already
circulating on streaming platforms,
often featured in suggested playlists,
with some generating substantial
streams
• Tools from third-party players such
as Boomy facilitate the creation and
upload of these Gen AI tracks on DSP
platforms
• The impact of this phenomenon, in
terms of volume of tracks and
streams, has yet to be quantified
• This raises questions about how
platforms should handle these
tracks (whether they should be
tagged for user identification and/or
removed)
• Managing this influx is challenging,
as streaming services now receive
about 1 million new songs each
week
Example of AI-generated songs on Spotify Boomy is a platform allowing the creation of
AI-generated music to be uploaded on DSPs
Boomy allows users to :
1. Create and edit songs
2. Release created music on DSPs
3. Use the musical work for :
• Non-commercial purposes in video,
livestreaming, and other songs
• Commercial purposes in podcasts and social
media and social media advertising
1
2
3
• In 2023, Boomy had created 14.4 million songs.
• The platform retains the copyright for all songs created, while
users receive an 80% share of the royalty distribution fees
Gen AI economic impact - Use cases
Automated complete musical output generation for
commercial distribution (e.g. “mood playlists”) 5
Music

57/116

58
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
2028 est. level of
adoption/maturity
Rise of tailored AI-generated music for social media content
Current level of
adoption/maturity
• Techno & quality of outputs: Current services
can produce quality musical content for such
purposes
• Usage/adoption: Mass use remains limited as
these tools are not fully integrated within major
social networks functionalities
Thanks to Generative AI, content creators can quickly create royalty-free music
perfectly suited to their YouTube videos for instance
CMO
Current
application
• Music generation for social media content using
Gen AI-powered tools is already underway
• Platforms are already investing in this technology
(e.g., TikTok, with the acquisition of Jukedeck in 2019)
or developing their own tools (e.g., Meta with AudioCraft)
• In addition to the current prompt-to-music system, Gen
AI will allow to provide instant, context-aware music for
user-generated content across all social media
platforms
• Platforms will continue to invest in and promote Gen AI
(copyright-free) music in their music libraries for
content creators
Example
A TikTok video with AI-generated music specifically tailored to the
video’s content
2028
potential
application
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
• Techno & quality of outputs: Improved
technology with increasingly higher-quality
outputs, with tools that are already user-friendly,
easy to use, and feature advanced UX
• Usage/adoption: Widespread use can be
expected with the integration of AI music
generation tools into major social networks
Low Mid High
Legend
(3)
Generalist players Specialized players Internal tools
Example of service providers
AudioCraft
by
Automated complete musical output generation for
background content (AV works, in-store, social media…) 7
Main economic impacts identified
High potential cannibalisation of a
portion of music creators’ social
media revenues
2 Revenue loss
Revenues driven by B2B subscription
fees and direct orders (brands,
content creators…)
Gen AI
providers’
revenues
3
Very high penetration rate of Gen AI
outputs in user-generated content
on social media
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact - Use cases Music

58/116

59
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Widespread adoption of Gen AI music for background content in
audiovisual works or public places
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Current tools can
already produce good quality music for such
purposes
• Usage/adoption: Penetration rates remain
limited so far
Generative AI has the potential to significantly impact background music, particularly
in tasks where high volumes and quick production times are key, much like traditional
music libraries. CMO
Current
application
• Gen AI tools are already used to generate background
scores for various projects and applications (e.g.,
advertising, sound systems in public places etc.) but
remain limited so far
• In addition to the use for background scores, Gen AI
could provide customizable and context-sensitive
background music services for a wide range of
multimedia
Automated complete musical output generation for
background content (AV works, in-store, social media…) 7
Example of service providers
Example
AI-generated jingle for a TV show
2028
potential
application
2028 est. level of
adoption/maturity
• Techno & quality of outputs: Improved
technology with increasingly higher-quality
outputs
• Usage/adoption: High potential for adoption by
B2B clients to reduce costs
Low Mid High
Legend
Main economic impacts identified
High cannibalisation rates:
replacement of human produced
“production music” for B2B use
2 Revenue loss
Gen AI tools’ revenues driven by B2B
subscription fees and direct orders
Gen AI
providers’
revenues
3
Very high penetration rate of Gen AI
outputs in the music library market
segment
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact - Use cases Music

59/116

60
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Example - Meta invests in AI to improve its music library and
enhance content creators experience
Source: Photo credit : Social Media Examiner
Meta’s free access music library for social media posts
Meta has been active in the library
music market since 2017:
• The Meta Sound Collection offers a
library of Meta-owned audio clips
available for free
In parallel, Meta invests in AI to
expand its collection of copyright-free
music content :
• In 2023, Meta launched
AudioCraft, an open-source AI
model allowing the generation of
high-quality, realistic audio and
music from text-based user inputs
Automated complete musical output generation for
background content (AV works, in-store, social media…) 7
Gen AI economic impact - Use cases Music

60/116

61
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Example of service providers
Moderate penetration of Gen AI music for core content in
audiovisual works
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Current services
can produce quality musical content, but they
are not always considered sufficient to fully
replace high-impact and high-budget
commissioned creations
• Usage/adoption: Use remains limited
• Techno & quality of outputs: Gen AI services will
rapidly be able to offer highly qualitative content
• Usage/adoption: Expected to become a more
widely used tool in audiovisual production for all
application types such as series, movies, video
games etc., except for high-budget projects
requiring the support of famous industry names
Generative AI enables advanced personalisation and real-time music creation. In
video games for instance, it can generate continuous music streams in a specific
style, adapting the sound to match in-game events.
Tech Company
Current
application
• Very limited applications today, with substantial
musical content for audiovisual works remaining mostly
commissioned compositions
• Gen AI is used to generate outputs replacing
commissioned works for core content in certain
audiovisual works (lower production budget)
• In addition, Gen AI could be used for more advanced
personalization and real-time music creation (e.g.,
video games )
Automated complete musical output generation for core
content in audiovisual works (e.g., video games main theme) 6
Example
AI-generated substantial musical soundtrack for a video game
production
2028
potential
application
2028 est. level of
adoption/maturity
Low Mid High
Legend
Main economic impacts identified
High potential cannibalisation of
revenues for audiovisual music
composers (less orders /
commissioned works)
2 Revenue loss
Revenues driven by B2B subscription
fees and direct orders mainly from
B2B producers
Gen AI
providers’
revenues
3
High penetration rate of Gen AI
outputs for background music in AV
works, for cost reduction
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact - Use cases Music

61/116

62
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approach and Methodology
What will be the economic impact of Generative AI in the Music
field by 2028?
Main Applications
What will be the economic impact of Generative AI in
Audiovisual by 2028?
2028 forecast
Economic
impact
in Music
creation

62/116

63
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Projected evolution of Gen AI music outputs market size | €Bn, 2023 - 2028
Fully Gen AI outputs in Music are expected to be worth c.€16Bn in 2028, doubling
on average each year
Source: PMP Strategy analysis
2023 2024 2025 2026 2027 2028
€1Bn
€7Bn
€11Bn
Note: In this market size calculation, no distinction is made whether the music outputs are copyrightable or not |
(1) 2023-2028 CAGR
+100%
avg. per
year(1)
Gen AI based music streaming
Gen AI music in traditional streaming
Gen AI music for B2B purposes
0.5
€Bn
4
€Bn
16
€Bn
1 Market size
Gen AI Music economic impact - Forecast

63/116

64
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Fully Gen AI music outputs market size | €Bn, 2028
This market will be mostly driven by Gen AI music on streaming platforms and
Gen AI music for B2B purposes
Source: PMP Strategy analysis
Note: In this market size calculation, no distinction is made whether the music outputs are copyrightable or not
€1Bn
€10Bn
€5Bn
Gen AI based
music streaming
Gen AI music
in traditional streaming
Gen AI music for B2B purposes
€16Bn
With the perfecting and democratization of prompt-to-”songs” (outputs), endconsumers are moving from simple users to music curators, thus questioning the very
notion of the creator. New “streaming” possibilities will likely emerge from this.
Tech company
Generative AI represents a potential opportunity for DSPs to generate royalty-free
tracks and integrate them into their playlists. This approach could significantly boost
their margins by drastically reducing copyright costs.
Tech company
Gen AI platforms like Boomy are disrupting the distribution system. They offer
prompter users the chance to directly distribute their creations (outputs) directly on
Spotify.
CMO
Generative AI has the potential to significantly impact background music,
particularly in tasks where high volumes and quick production times are key, much
like traditional music libraries.
Tech company
1 Market size
Gen AI Music economic impact - Forecast

64/116

65
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Market saturation will be significant on the music library segment (c.60%),
with B2B clients looking to reduce costs
Source: PMP Strategy analysis
Gen AI based music
curation and streaming 100%
Gen AI music in
traditional streaming
platforms
20%
Gen AI music for B2B
purposes
(audiovisual, social
media, public places
sound system, brands)
57%
• New services allowing both music listening and curation/creation (see
associated use case)
• 2 potential scenarios: (i) Tools like Suno and Udio become new players in the
music streaming industry, (ii) Existing music streaming platforms integrate
and monetize these new Gen AI features
• Penetration of Gen AI outputs in DSPs catalogues, generated by third parties or
by the streaming platforms themselves (see associated use case)
• High potential in mood music playlists (e.g., “Morning Coffee”, “Beast Mode” on
Spotify) where end-users adopt a more passive listening
• Use of fully Gen AI music for background music in audiovisual works,
advertising, social media, in-store sonorization (see associated use cases)
• High adoption rates fostered by B2B clients looking for costs reductions and the
unlimited potential of these Gen AI outputs for B2B consumers (brands,
audiovisual professionals, content creators, etc…)
Market segments Fully Gen AI penetration rate in 2028(1) Use cases supporting the penetration of Gen AI outputs
Note:
(1) Weighed penetration rates of subcategories analysed
1 Market size
Gen AI Music economic impact - Forecast

65/116

66
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
99%
1%
5%
94%
<1%
18%
80%
3%
€36 Bn
€43 Bn
Music streaming B2C(1) revenues, generated by human created works vs. Gen AI outputs | €Bn, 2023 - 2028
A Gen AI boost is expected on the music streaming segment, due to new usage and
functionalities which will be monetized by traditional or new players
Source: PMP Strategy analysis
€57 Bn
Gen AI outputs penetrate the market in the
traditional music streaming platforms and
in new Gen AI based streaming services
In 2028, Gen AI outputs could create an
estimated additional €1.4Bn in the market,
but the major part (c.€10Bn) of the revenue
generated will be a substitution of humancreated works’ revenues
In the current market, 99% of the € 36Bn
music streaming market is generated by
human creations
Gen AI boost –
additional value (2)
Gen AI music in
streaming – replacing
human music
Human-created
works in streaming
Human-created
works in streaming
Gen AI music in
streaming
2023 2025 2028
Note: (1) Music streaming platforms B2B revenues (brands, in-store sonorization…) have been excluded of this analysis (included in the B2B
segment of the market size calculation) | (2) Gen AI based music platforms (new Gen AI platforms or new offers of traditional DSPs)
1 Market size
Gen AI Music economic impact - Forecast

66/116

67
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Revenues for Music creators with and without the impact of Generative AI | €Bn and %, 2023 - 2028
Under current conditions, this market penetration by Gen AI outputs could put 24%
of Music creators’ revenues at risk by 2028
Source: PMP Strategy analysis, CISAC global collections report
Note: In this analysis, creators' revenues are represented by CMOs collections
Share of creators’ revenue at risk due to Gen AI substitution
Revenues for creators with the impact of Gen AI
2028 total
cannibalisation
% on total revenues
€4Bn
24%
∑= €10
Bn
10
16
2023 2024 2025 2026 2027 2028
’23-’28
cumulated loss
Revenue 2 loss
Gen AI Music economic impact - Forecast

67/116

68
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The potential impact will be strong on Digital collections (up to 30% cannibalisation),
TV & Radio and Background (c. 22% of collections)
Source: PMP Strategy analysis
Note: (1) Cannibalisation rates estimated based on interviews and workshops with CMOs and industry experts
Gen AI cannibalisation rate in 2028(1) Use cases explaining cannibalisation levels Music creators’ revenue
streams
Digital
TV & Radio(2)
Live & Background
CD & Video
Other streams
30%
22%
22%
21%
• Gen AI music in social media content, audiovisual streaming (AV works’
background music), music streaming (mainly in mood music) and video games
• Gen AI music for background content in audiovisual works (mainly in lower
production budget works)
• Gen AI music for the sonorisation of public places (stores, malls, restaurants; using
personalised and unlimited Gen AI outputs to reduce costs)
• Gen AI music in video games (use case with high level of adoption, due to new
unlimited tracks for games – by moods, styles – enabled by Gen AI)
Marginal to no economic impact
Revenue 2 loss
Gen AI Music economic impact - Forecast

68/116

69
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music creators’ revenue share in the music streaming market | €Bn, 2023 - 2028
In a growing music streaming market, creators’ share will thus decrease further
due to Gen AI (-1.5pts)
Source: PMP Strategy analysis, CISAC global collections report,
• In 2023, the share of creators' revenue
in the streaming market amounts to
approximately 8%
• In 2028, this share could decrease to
c.6%, on a significantly higher market
• This dilution could represent a loss of
c.€0.9Bn for creators in 2028 and a
cumulated loss of c.€2.3Bn in the five
coming years
Note: (1) Music streaming platforms B2B revenues (brands, in-store sonorization…) have been excluded of this analysis
(included in the B2B segment in the market size calculation
Revenue for creators
without Gen AI
Revenue for creators with
Gen AI
2023 2024 2025 2026 2027 2028
5%
10%
8.2% 8.2%
8.0%
8.2%
7.5%
8.2%
7.2%
8.2%
6.8%
8.2%
6.4%
c.0.9Bn in revenue loss
for music creators in
streaming only
Revenue 2 loss
Gen AI economic impact - Forecast Music

69/116

70
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music Gen AI providers’ revenues | €Bn, 2023 - 2028
Gen AI providers’ revenues in Music could reach c.€4Bn in 2028, doubling on average
each year from 2023 to 2028
Source: PMP Strategy analysis
2023 2024 2025 2026 2027
€1Bn
€2Bn
2028
€0.3Bn
€1Bn
€2Bn
Mass public tools for complete outputs’ generation
Gen AI softwares for assistance in the creative process
c.0.1
€Bn
c4
€Bn
c.0.6
€Bn
Note: (1) 2023-2028 CAGR
Gen AI services 3 revenues
Gen AI Music economic impact - Forecast
+100%
avg. per
year(1)

70/116

71
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries Source: PMP Strategy
Gen AI outputs in Music will be worth a cumulative €40Bn over the next five
years, rising to an annual value of €16Bn in 2028
By 2028, Gen AI music will account for around 20% of traditional music streaming
platforms’ revenues and around 60% of music libraries revenues
Market
size 1 €16Bn
Estimated market value of Gen
AI outputs in Music in 2028
Under current conditions, this market penetration by Gen AI outputs could put 24% of
Music creators’ revenues at risk in 2028
This represents a cumulative loss of €10Bn over the next 5 years, and an
annual loss of €4Bn in 2028
Revenue
loss 2 €4Bn|24%
Creators' revenues at risk in 2028
compared to a no Gen AI situation
Gen AI
services’
revenues
2 €4Bn
Estimated revenues of Gen AI
Music services in 2028
Gen AI services are projected to generate exponential revenue growth, reaching an
estimated €4Bn in 2028, with a cumulative total of €8Bn over 5 years
Study key takeaways – Music
Gen AI economic impact - Forecast Music

71/116

72
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approach and Methodology
What will be the economic impact of Generative AI in the Music
field by 2028?
What will be the economic impact of Generative AI in
Audiovisual by 2028?
Main Applications
2028 forecast
Economic
impact in
Audiovisual
creation

72/116

73
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Identification of Gen AI main applications in the audiovisual creation process
Source: Experts’ interviews, specialized press
Ai-assisted video editing, mastering, voice
removing, visual effects…
Dubbing-subtitling
Automated translation and adaptation
AI-assisted storyboard
generation, scene visualization
Creative inspiration / Research and planning Creation / Execution Post-production
Main use case
AI-assisted or automated complete
audiovisual outputs creation/composition AI-assisted post-production AI-assisted video & audiovisual workconceptualization/idea generation
7
Automated complete audiovisual output generation
for higher production budget works (series, films…)
Automated complete audiovisual output generation
for entertainment purpose (mainly short form) 3
1 8
Screenwriting
Automated screenwriting 2
Automated complete audiovisual output generation
for lower production budget works (ads, children’s
cartoons, ..)
4
AI-assisted or automated complete audiovisual
works generation for social media content
(short form, mainly user generated content)
5
6
Gen AI economic impact – Use cases Audiovisual

73/116

74
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Prioisation of use cases based on expected impact on creators’ revenues and adoption probability – Matrix Analysis
Prioisation of Gen AI use cases in the Audiovisual field
Source: PMP Strategy Analysis
Very Low
Very High
Adoption probability
Very Low
AI-assisted Video editing, Mastering, Voice
removing, Visual Effects… 8
AI-assisted Storyboard
generation, Scene visualization 1
Automated complete audiovisual output generation
for higher production budget works (series, films…) 5
Automated complete audiovisual output generation
for entertainment purpose (mainly short form) 3
Automated translation and adaptation 7
AI-assisted or automated complete audiovisual works
generation for social media content
(short form, mainly user generated content)
6 Main use cases
Mid to High adoption
Low impact
High impact
Low adoption
Automated screenwriting 2
Automated complete audiovisual output generation for lower
production budget works (ads, children’s cartoons, ..) 4
Impact on creators’ revenues
Very High
Gen AI economic impact – Use cases
Legend: use cases impact on creators’ revenues
- Marginal - Moderate - High to very high
Audiovisual

74/116

75
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Example of service providers
Widespread adoption of Gen AI tools for video content generation on
social media
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Currently, AI
services cannot produce complex and highly
customized visuals using the prompt-to-video
method
• Currently, adoption is low, as the technology is
still developing
• With expected strong improvement in
technology capacities, by 2028, the adoption is
expected to be very high, with AI-assisted tools
becoming standard in the content creation
process, particularly for creators who need to
produce high-quality, visually rich content on
tight deadlines
Current
application
• Generation of illustrative, moderate quality videos that
support and enhance various types of content
• Automating tasks like adding relevant visuals,
animations, and effects, allowing them to produce
engaging content more efficiently
• Generation of high quality, longer video sequences with
minimal input
Example
AI-assisted or automated complete audiovisual
works generation for social media content
(short form, mainly user generated content)
6
Illustrative videos supporting a history-themed video on YouTube
Make-a-Video by
2028
potential
application
2028 est. level of
adoption/maturity
Low Mid High
Legend
Gen AI will democratize creation and increase user-generated audiovisual
content. If technology allows it, we'll have tools enabling to create affordable,
high-volume short videos for social networks which will flood the market.
Audiovisual institution
Main economic impacts identified
High cannibalisation of creators’
revenues from traditional video
production services
2 Revenue loss
Revenues driven by subscription
fees from content creators and other
social media users
Gen AI
providers’
revenues
3
High penetration rate and market
growth with the use of Gen AI video
mainly for content creators
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact – Use cases Audiovisual

75/116

76
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Gen AI automated subtitling for a documentary production
Improving technology for the automation of translations
and adaptations
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Currently, AI
services can produce complex and highly
customized visuals using the prompt-to-video
method
• Currently, adoption is low for complex and
high-quality audiovisual works but is high for
lower value content
• By 2028, adoption is expected to be very high,
with automated dubbing and subtitling
becoming more qualitative and therefore a
standard practices in the industry, particularly
for streaming platforms and global content
distributors.
If the technology improves, Gen AI could be used to produce high-quality
content. However, as of now, automated subtitling and dubbing are still limited
to low added-value applications.
Dubbing and Subtitling Agency
Current
application
• Automating speech translation, subtitles
synchronisation and voice dubbing generation
• Application is currently focused on non-substantial
video due to technology capacity
• High-quality real-time translations and lip-syncing
that are indistinguishable from human performance
Example
7 Automated translating/adapting
Example of service providers
2028
potential
application
2028 est. level of
adoption/maturity
Low Mid High
Legend
Main economic impacts identified
High cannibalisation of revenues for
DB/ST authors and less orders for
DB/ST agencies
2 Revenue loss
Revenues driven by subscription
fees from audiovisual producers for
DB/ST works
Gen AI
providers’
revenues
3
High penetration rate of Gen AI
DB/ST; potential shrinking of the
overall market (cost reduction)
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact – Use cases Audiovisual

76/116

77
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Fully Gen AI automated screenwriting for a TV soap opera
Generative AI as an assistant and/or a substitution for
screenwriting in audiovisual works
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: easy to use tools
• Currently, adoption is still pretty low as tools are
mostly used for assistance / idea generation or
rewriting
• By 2028, adoption is expected to be very high
especially on lower production value segments,
with AI tools allowing to create ever more
complex stories based on specified criteria
The main concerns about the impact of Gen AI often arise from screenwriters among
our members: the profession is likely to be heavily impacted in the coming years.
Audiovisual CMO
Current
application
• Current tools allow to automate / facilitate a number of
tasks related to scriptwriting: scenario analysis,
research, rewriting, …
• The progress of tools will allow to generate more quality
scripts and automate the full generation of scenarios
for certain contents
Example
Example of service providers
2028
potential
application
2028 est. level of
adoption/maturity
Low Mid High
Legend
Main economic impacts identified
High cannibalisation of revenues
for screenwriters as Gen AI scripts
become more cost-effective
2 Revenue loss
Revenues driven by subscription
fees from audiovisual producers for
screenwriting works
Gen AI
providers’
revenues
3
High penetration rate of Gen AI
scripts; potential shrinking of the
overall market (cost reduction)
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact – Use cases
2 Automated screenwriting
Audiovisual

77/116

78
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Example of service providers
Rise of Gen AI content for lower budget audiovisual productions,
fostered by producers’ willingness to gain efficiency
Source: Experts’ interviews, Specialized press, PMPS Scenario Planning
Current level of
adoption/maturity
• Techno & quality of outputs: Currently, AI
services cannot produce quality audiovisual
content
• Currently, adoption is low, due to the poor
quality of AI-automated audiovisual outputs
generation
• Mid- to high adoption, depending on the
evolution of the quality of the audiovisual
content/output
Current
application
• First creations of entire audiovisual outputs in animated
fiction films & series
• Ads or Music clips generation, with enhanced
possibilities but mid production quality
• High-quality audiovisual content that is indistinguishable
from human-made productions
• Potential seamless integration into some audiovisual
production sectors (soap operas, advertising, music clips,
animated works…)
Automated complete audiovisual output
generation for lower production budget
works (ads, soap opera...)
4
Example
Extract of Qianqiu Shining, China Media Group AI-generated animated series
in 2022
Adoption depends on the technology's capabilities. Currently, producing a
high-quality audiovisual work from start to finish, especially a two-hour film, is
beyond reach, while it is already feasible for a short-animated film.
Audiovisual institution
2028
potential
application
2028 est. level of
adoption/maturity
Low Mid High
Legend
Main economic impacts identified
Very high impact on audiovisual
creators/authors due to productions’
budget decrease
2 Revenue loss
Revenues driven by B2B subscription
fees from
audiovisual production companies
Gen AI
providers’
revenues
3
High penetration rate of Gen AI;
shrinking of the traditional market
(cost reduction)
1
Market size /
Gen AI
penetration
Expected impact on
creators’ revenues
Gen AI economic impact – Use cases Audiovisual

78/116

79
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Approach and Methodology
What will be the economic impact of Generative AI in the Music
field by 2028?
What will be the economic impact of Generative AI in
Audiovisual by 2028?
Main Applications
2028 forecast
Economic
impact in
Audiovisual
creation

79/116

80
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Projected evolution of Gen AI audiovisual outputs market size | €Bn, 2023 - 2028
Fully Gen AI audiovisual outputs are expected to be worth c.€48Bn in 2028, with
an average growth of c.85% each year
Source: PMP Strategy analysis
2023 2024 2025 2026 2027 2028
€6Bn
€22Bn
€34Bn
Note: In this market size calculation, no distinction is made whether the Gen AI outputs are copyrightable or not, and only Fully Gen AI
audiovisual/video outputs are considered |
(1) 2023-2028 CAGR
AVOD (incl. Youtube)
Video streaming (SVOD & downloads)
TV broadcasters
Social media
13
€Bn
48
€Bn
2
€Bn
+85%
avg. per
year(1)
Gen AI economic impact - Forecast
1 Market size
Audiovisual

80/116

81
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Fully Gen AI audiovisual outputs’ market size | €Bn, 2028
Source: PMP Strategy analysis
Note: In this market size calculation, no distinction is made whether the Gen AI outputs are copyrightable or not | (1) includes SVOD, FAST,
Pay-per-view, EST (downloads)
€11Bn
€3Bn
€5Bn
€30Bn
TV broadcasters
SVOD
AVOD (incl. Youtube)
Social media
€48Bn
Some specific YouTube creators now have the potential to produce fully Gen AI
videos, enabling them to deliver high-quality, engaging content on complex topics
with minimal effort. For instance, a channel dedicated to science education can
leverage Gen AI to generate fully automated content.
Audiovisual institution
There is a risk that some audiovisual production with lower budget might be replaced
in the future by Gen AI outputs, for example some soap operas, low-budget
advertising & clips, or animated content for kids… Producers of these content will be
looking to reduce costs by leveraging Gen AI tools.
CMO
Gen AI will democratize creation and increase user-generated audiovisual content. If
technology allows it, we'll have tools enabling to creating affordable, high-volume
short videos for social networks which will flood the market.
Audiovisual institution
(1)
The market will be mostly driven by the penetration of Gen AI outputs on social media
and in lower production value TV programmes
Gen AI economic impact - Forecast
1 Market size
Audiovisual

81/116

82
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Market saturation by Gen AI complete audiovisual outputs will remain more
limited than for Music
Source: PMP Strategy analysis
4%
2%
TV Broadcasters
SVOD
AVOD 8%
• Leveraging of Gen AI content to produce lower budget and more personalized
programmes (lower budget TV shows, advertisings, kids animated, …) allowing
to keep up with audience demand at reduced costs
• Penetration of Gen AI audiovisual outputs in SVOD platforms catalogues
• Use of Gen AI tools to produce audiovisual outputs for short and viral content
on AVOD platforms (mainly YouTube)
Market segments Fully Gen AI penetration rate in 2028(1) Use cases supporting the penetration of Gen AI outputs
Note:
(1) Weighed penetration rates of subcategories analysed
Social media 13%
• Use of Gen AI tools to automate the generation of short form videos on social
media, and produce engaging and personalized content more efficiently by
adding relevant visuals, animations, and effects
1 Market size
Gen AI economic impact - Forecast Audiovisual

82/116

83
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Revenues for Audiovisual creators with and without the impact of Gen AI outputs | €Bn and %, 2023 - 2028
The use of Gen AI tools to automate tasks in the production process could put 21%
of audiovisual creators’ revenue at risk by 2028
Source: PMP Strategy analysis, CISAC global collections report
2028 total
cannibalisation
% on total revenues
€4.5Bn
21%
∑= €12
Bn
Note: In this analysis, revenues include both CMOs collections and other revenue streams (upfront payments)
10
22
2023 2024 2025 2026 2027 2028
Share of creators’ revenue at risk (due to substitution of complete Gen AI outputs and use of Gen AI tools in the production process)
Revenues for creators with the impact of Gen AI
’23-’28
cumulated loss
Gen AI economic impact - Forecast
Revenue 2 loss
Audiovisual

83/116

84
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The potential impact will be particularly strong for Translators and Adapters
(c.56% of cannibalisation rate)
Source: PMP Strategy analysis
Note: (1) Cannibalisation rates estimated based on interviews and workshops with CMOs and industry experts
Directors & other
co-authors
Translators /
Adapters
15%
20%
56%
• Widespread use of Gen AI tools to automate directors and other co-authors’
tasks, fostered by producers’ willingness to gain efficiency and reduce costs
• Complete audiovisual outputs replacing human-created works on certain
categories of audiovisual content
• Use of Gen AI tools for automatic translation and adaptation, with outputs
increasingly closer to human work at a decreasing cost
Gen AI cannibalisation rate in 2028(1) Use cases explaining cannibalisation levels Audiovisual creators /
authors’ categories
Screenwriters • Use of Gen AI screenwriting assistance tools, supporting authors in their work but
also pushing producers to reduce the budget spend for screenwriting
Revenue 2 loss
Gen AI economic impact - Forecast
84
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual

84/116

85
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual Gen AI providers’ revenues | €Bn, 2023 - 2028
In Audiovisual, Generative AI providers’ revenues could reach €5Bn in 2028,
driven by Gen AI prompt-to-outputs tools
Source: PMP Strategy market model, Experts’ interviews
2023 2024 2025 2026 2027
67%
33%
2028
€0.5Bn
€2Bn
€4Bn
Audiovisual prompt to outputs services market size
Gen AI-assisted video/audiovisual creation softwares
c.0.2
€Bn
c.5
€Bn
c.1
€Bn
Gen AI services 3 revenues
+85%
avg. per
year(1)
Note: (1) 2023-2028 CAGR
Gen AI economic impact - Forecast Audiovisual

85/116

86
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries Source: PMP Strategy
AI-generated complete Audiovisual outputs are expected to be worth c. €48Bn in
2028.
Audiovisual outputs generation for social media and TV will account for the lion’s
share of the market.
Market
size 1
€48Bn
Estimated market value of Gen AI
outputs in Audiovisual in 2028
The widespread use of Gen AI tools throughout the production process of
audiovisual works could put 21% of creators' revenue at risk by 2028.
This represents a cumulative loss of €12Bn over the next 5 years, and an annual
loss of €4.5Bn in 2028.
Revenue
loss 2
Gen AI
services’
revenues
2
€5Bn
Estimated revenues of Gen AI
Audiovisual services in 2028
Gen AI services in Audiovisual (both mass public and professional tools/softwares)
are projected to generate exponential revenue growth, reaching an estimated €5Bn
in 2028, with a cumulative total of €13Bn over 5 years.
Study key takeaways – Audiovisual
€4.5Bn|21%
Audiovisual creators' revenues
at risk in 2028 (compared to a
no Gen AI situation)
Gen AI Audiovisual economic impact - Forecast

86/116

87
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Detailed methodology and assumptions – Music
Glossary
Detailed list of interviews conducted
PMP Strategy presentation
Appendix

87/116

88
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Market segmentation and impact of Gen AI use cases done by crossing 3 dimensions
Use cases analysis has allowed to identify the Music market segments most
likely to be impacted by Gen AI in the next 5 years
Source: PMP Strategy methodology
Legend: expected impact of Gen AI outputs on the market segment (market penetration of Gen AI outputs)
- Marginal to low - Moderate - High to very high
Production budget of the works
Higher budget works
Lower budget works
Type of musical artwork
Commercial music
Mood music
Background music/sound effects
(for social media, audiovisual, public places)
Core music in audiovisual content
Music works distribution channel
Digital – current streaming platforms
Digital – New AI-based streaming
platforms
Digital – other downloads
Physical sales (CDs, vinyls...)
Music libraries
Direct commissioned works
B2C / B2B2C:
B2B – Music for audiovisual, social
media, in-store sonorization, brands…:
Concerts/Festivals/Music shows
(Live audience)
1 Market size
Appendix – Detailed methodology and assumptions Music

88/116

89
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
For each of these segments, a Gen AI penetration rate has been estimated based
on use cases expected adoption and impact (from low to very high)
Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews
2023 market size and 2028 forecasts
on segments impacted by Gen AI Fully Gen AI outputs Penetration rate in 2028 2028 Fully Gen AI music outputs’ market size
1 Market size
x =
‘23 market size ’28 forecasts
c. €16Bn
New Gen AI
music streaming
platforms(1)
Music streaming
platforms
<0.1Bn
€36Bn
€5Bn
1.4Bn
€55Bn
€7.8Bn
• New Gen AI based streaming platforms allowing listeners/users to
curate and listen to Gen AI music and/or new offers in current
streaming platforms
• Commercial music
• Mood music
• Licensed music – commercial/preexisting tracks
• Licensed music – background/sound effects
• “Royalty free” music (buy out)
100%
20%
57%
€1.4Bn
€10Bn
€4.5Bn
Total Gen AI market size for musical outputs
• We consider here the “public price” of music for B2B clients buying / commissioning musical works for audiovisual content, public places sonorization, etc.
• The impact on the value driven by the diffusion of this music in audiovisual works (by SVOD platforms, TV/Radio broadcasters) is considered as part of question 2/
• The market size calculation includes royalty-free music content
Note: (1) Does not include all AI-based music services, but only new platforms on which end-users/subscribers are both music listeners
and creators (Udio, Suno) | (2) Also includes music libraries from majors (e.g. Universal Music Production), and from DSPs
B2C/B2BC2 B2B
Music libraries(1)
-
for audiovisual,
social media,
public places
sonorization,
brands
Weighed penetration rates considering :
Appendix – Detailed methodology and assumptions
89
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music

89/116

90
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Note: (1) Segmentation of this category crossing the type of audiovisual content on which it is played and the importance of the music
in the AV work
Music market segmentation on segments impacted by AI generated outputs
Revenue cannibalisation for Music creators has been calculated based on a
segmentation of the global CMOs collections
Source: PMP Strategy market model, Experts’ interviews
Revenue
loss 2
• Segmentation approach to
measure AI outputs’
cannibalisation on creators’
revenues:
– Breakdown of CISAC 2023
global Music collections by
categories (e.g Digital) and
subsegments (e.g Music
streaming platforms, SVOD
platforms, Social networks, etc.)
– 2028 collections forecast for
each subsegment based on
historical growth rates and
future market trends
– For each subsegment:
estimated cannibalisation
rate in 2028 based on use
cases and market estimates
conducted as part of question
1/
Key methodology insights
Digital
CISAC Global collections
Collections segmentation by use Rationale
• Exclusion from the calculation (blurred boxes) of the collections’ subsegments where the human created works
are unlikely to be replaced by AI:
• E.g. : it is estimated that the music used for live shows will remain predominantly human-created, as the
audience always associates itself with a musician when attending a live event
TV & Radio(1) CD & Video synchronisation Others Live &
Background
Music streaming
AV streaming
Social networks
Video Games,
Download &
other digital
Original music of
high production
value AV work
Background music
for high production
value AV works
Original music of low
production value AV
work
Background music
of low production
value AV wors
Live audience
performances
Background music
Music as
main focus in
public places
Other sources
CD, Vinyls
DVD, Blue rays
Video games
Appendix – Detailed methodology and assumptions Music

90/116

91
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
CISAC Music collections evolution by music usage in current market evolutions | €Bn, 2018-2028
Based on historical growth and future market trends, CISAC collections for Music
are expected to reach c. €15.8Bn by 2028
Source: PMP Strategy market model, Experts’ interviews
Forecasts
0%
19%
4%
2018 2019 2020 2021 2022
39%
29%
26%
3%
0%
3%
2023 2024 2025 2026 2027
44%
22%
27%
3%
39%
1%
30%
3%
2028
8.5
8%
8.2 8.5
10.8
11.8
8.4
13.2
14.0
14.9
15.8
12.4
Digital TV & Radio Live & Background CD & Video Synchronisation Others
YoY growth (%) -1% -2% +4% +28% +9% +6% +6% +6% +6% +6%
CAGR
18-22 22-24
+6% +7%
+26% +10%
+2% -2%
24-28
+6%
+9%
+1%
-1% +14% +7%
-15% +6% +2%
+13% +14% +10%
-2% +13% +6%
Post-Covid
recovery
15.8
€Bn
Note: Forecasts CAGR based on historical collections and upcoming trends
Revenue
loss 2
Appendix – Detailed methodology and assumptions Music

91/116

92
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
To estimate the revenue loss for Music creators, key assumptions have been made
on Gen AI cannibalisation rates by collections subsegments
Source: PMP Strategy market model, CISAC collections, Specialized press, Experts’ interviews
Revenue
loss 2
‘23 collections ’28 forecasts
c. €3.8Bn (24%) Total revenue loss for
music artists in 2028
x =
Revenue sub-streams (1), 2023 collections and 2028
forecasts
2028 revenue loss
for creators
Digital €4.5Bn
• Music streaming
• AV streaming
• Social media
• Video Games
€6.9Bn 30% €2Bn
TV & Radio(1) €3.4Bn
• Original music for higher budget AV works/ music tracks on radio
• Background music for higher budget AV works
• Original music for lower budget AV works
• Background music for lower budget AV works
€3.5Bn 22% €0.7Bn
Live &
Background €3.1Bn
• Live audience performances
• Music as main focus in a live audience (clubs…)
• Background music (sonorization of public places)
€4.2Bn 21% €0.8Bn
CD & Video €0.4Bn
• CD, Vinyls
• DVD, Blue-rays
• Video games
€0.4Bn 21% €0.1Bn
Weighed cannibalisation rates considering :
Gen AI outputs 2028 cannibalisation rates
Note: (1) AV works in the cannibalisation rates consider both TV & radio works
Other substreams €0.3Bn €0.7Bn
Total collections €11.8Bn €15.8Bn
Marginal to no impact
Appendix – Detailed methodology and assumptions Music

92/116

93
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Music Gen AI tools providers’ revenues – 2023-2028, €Bn
Gen AI tools providers have been segmented in 2 categories to assess their
revenues
Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights
1. Mass public
tools
2. Gen AI
tools/software as
an assistant
Number of Gen AI based
services in 2028
Average Number of users in 2028
and % of paying subscriptions
Average annual price of the
subscription
€1.4Bn(1)
2028 associated players
total revenues
4
60m
(10%) €60
x x =
Audio/Music software market size
2023 2028
€2.3Bn €4.2Bn(2)
2028 Generative AI tools penetration rate
€2.3Bn
2028 associated players
= total revenues
60% / High
x
Note: (1) Calculated in question 1, representing new AI based streaming platforms (and described as the use case: end-user as curators) |
(2) High growth rate as new services (providing both prompt-to-outputs and assistance in the creation – such as Boomy and Beatoven)
will lead to revenue growth on the market
Gen AI services 3 revenues
Appendix – Detailed methodology and assumptions Music

93/116

94
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Detailed methodology and assumptions – Audiovisual
Glossary
Detailed list of interviews conducted
PMP Strategy presentation
Appendix

94/116

95
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual market segmentation(1) and impact of Gen AI use cases done by crossing 3 dimensions
Use cases analysis has allowed to identify the Audiovisual market segments
most likely to be impacted by Gen AI in the next 5 years
Source: PMP Strategy methodology
- Marginal - Moderate - High to very high
Artistic value of the works / Nature of
the programmes
Type of audiovisual content Audiovisual works distribution
channels/media
Cinema
TV broadcasters
Radio broadcasters
Digital - AVOD
Digital - SVOD
Digital - Social Media
Physical sales (DVD, Blue-ray, …)
Complete Audiovisual content :
Films / Series
TV magazines/TV news/weather/
sports event
Soap Opera
Music clips
Short forms / User-generated content
Higher budget works
Lower budget works
Legend: expected impact of Gen AI works on the market segment (market penetration of Gen AI works)
Note: (1) Audiovisual market excluding audio podcast and video games
Flow programmes*
Stock programmes*
Animated works/Kids shows
* Stock programs can be rebroadcast (e.g., fiction,
documentaries), while flow programs are usually aired
once (e.g., news, sports, games)
Advertising
1 Market size
Appendix – Detailed methodology and assumptions Audiovisual

95/116

96
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
For each of these segments, a Gen AI penetration rate has been estimated based on
use cases expected impact (from low to very high) – Focus on complete AV outputs
Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews
2023 market size and 2028 forecasts
on segments impacted by Gen AI Fully Gen AI artworks Penetration rate in 2028 2028 Fully Gen AI AV works’ market size
1 Market size
x =
‘23 market size ’28 forecasts
c. €48Bn Total Gen AI market size for
complete audiovisual productions
Disclaimer: Market size including royalty-free audiovisual content (mainly videos)
Weighed penetration rate, including(1):
TV broadcasters €327Bn €298Bn
• Flow programmes of which news, weather, feature stories…)
• Flow programmes of which sports events, games…
• Ads & clips (lower budget works)
• Stock programmes – of which films, series…
• Stock programmes – of which kids animated works…
4% €11Bn
Digital - SVOD €107Bn €161Bn 2% €3Bn
• Higher budget AV works (films for cinema…)
• Lower budget complete AV works (daily soap opera, reality TV…)
• Animated works/Kids' content
Digital - AVOD €37Bn €60Bn 8% €4.5Bn
• Music/covers videos
• Educational/tutorial
• Gaming videos
• Corporate videos
• Others (incl. vlog, comedy)
Digital - Social
Media €186Bn €230Bn 13% €30Bn • Short form videos (Reels, TikTok)
• Long-format videos
Appendix – Detailed methodology and assumptions Audiovisual

96/116

97
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual collections segmentation and identification of segments on which AI impact is marginal
Revenue cannibalisation for creators has been calculated based on a
segmentation of Audiovisual global collections
Source: PMP Strategy market model, Experts’ interviews
1. Isolation of collections for
translators/adapters and
screenwriters, and calculation of the
revenue loss of these categories
based on cannibalisation rates
assumptions for each AV works
represented
2. For the other works (complete
audiovisual outputs): Bottom up &
Top-down approach to calculate the
potential revenue loss for complete
audiovisual outputs :
– Breakdown of CISAC 2023 global
audiovisual collections by
categories and subsegments
– 2028 collections forecast for each
subsegment based on historical
growth rates and future market
trends
– Estimation of Gen AI outputs’ works
cannibalisation rate in 2028 based
on use cases and market
estimates conducted as part of
question 1.
Key methodology insights
CISAC global collections
(for directors & other co-authors)
Collections segmentation by use
• Exclusion (or marginal impact on the total) from the calculation of the collections’ subsegments where the AI generated works
are unlikely to be cannibalized by Gen AI (blurred boxes):
• E.g., (i) Marginal impact on radio collections as few collections from radio components likely to be impacted by AI (audio
podcasts) (ii) Private copying works may shrink due to shrinking of global revenues but collections will still be from
Rationale
human creations
Private
Copying Digital
SVOD
AVOD
Social Media
TV & Radio
Radio
TV
Live &
Background
Background
Collections
translators/adapters
Others
Educational use
Mechanical
reproduction
Rental, Public
Lending
Others
Collections
screenwriters
Revenue
loss 2
Appendix – Detailed methodology and assumptions Audiovisual

97/116

98
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
CISAC audiovisual’s collections evolution by categories/revenue streams in current market evolutions | €m, 2018-2028
Based on historical growth and future market trends, CISAC collections for
Audiovisual are expected to reach 814m by 2028, with a 24-28 CAGR of c.3%
Source: PMP Strategy market model, Experts’ interviews
Forecasts
2%
0% 2%
45%
4%
2018 2019 2020 2021 2022
35%
6%
6% 2%
2% 0%
45%
4%
2023 2024 2025 2026 2027
31%
6%
10%
2%
38%
1%
7%
2%
45%
4%
1%
605 597 626 608
646
2028
710 732 757 784
814
690
TV & Radio
Private Copying
Digital
Live & Background
Educational use
Others
For Screewriters
For Translators & adapters
YoY growth (%) -1% +5% -3% +6% +7% +3% +3% +3% +4% +4%
CAGR
18-22 22-24
+2% +5%
0% +3%
-1% +2%
24-28
+3%
+1%
+2%
+51% +17% +15%
-0 +12% +4%
+3% +3% +6%
-1% +2% +2%
Revenue
loss 2
+2% +5% +3%
+2% +5% +3%
For Directors & other authors contributing
to the making of audiovisual works
Appendix – Detailed methodology and assumptions Audiovisual

98/116

99
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The revenue loss for directors and other authors contributing to the making of the AV
works has been estimated by applying 2028 estimated cannibalisation rates
Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews
x =
‘23 collections ’28 forecasts
Revenue sub-streams (1), 2023 collections and 2028
forecasts Gen AI works 2028 cannibalisation rates
2028 revenue loss
for creators
€240m €252m
c. €19m(1) (5%)
Total revenue loss for directors & other
authors contributing to the making of
audiovisual works
TV & Radio
• Flow programmes of which news, weather, feature stories…
• Flow programmes of which sports events, games…
• Ads & clips (lower budget)
• Stock programmes – of which films, series…
• Stock programmes – of which kids animated works
4% €9m
Digital €41m €82m
• AVOD: (weighed penetration rate from market)
• SVOD & digital TV
• Social media
€4m
€12m €15m Live &
Background
• Background audiovisual works
(considered as highly at risk as Gen AI will enable to broadcast
unlimited content in public places, stores…)
30% €5m
Note: (1) Includes cannibalisation for educational use (potential replacement for tutorials, webinars…) resulting in a €1m revenue loss
Weighed penetration rate, including(1):
2
Revenue
loss
€59m €66m Other substreams Marginal to no impact
Total collections €352m €415m
5%
Appendix – Detailed methodology and assumptions Audiovisual

99/116

100
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
The revenue loss for translators/adapters and screenwriters have been estimated
based on a forecast of 2028 revenues and the application of cannibalisation rates
Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews
x =
‘23 collections ’28 forecasts
c. €13m
Translation /
Adaptation €28m €33m • cannibalisation rate equivalent to the weighed penetration rates from
market size calculations 41% €13m
Screenwriting €311m €366m • cannibalisation rate equivalent to the weighed penetration rates from
market size calculations 19% €70m
Total revenue loss
for translators/adapters
Total revenue loss for screenwriters c. €70m
’23 revenue sub-streams and ’28 forecasts Gen AI works 2028 cannibalisation rates
2028 revenue loss
for creators
c. €0.1 Bn (12%) Total revenue loss in 2028 for
audiovisual creators
2
Revenue
loss
Total collections
(incl. the three
categories)
€690m €814m
Appendix – Detailed methodology and assumptions Audiovisual

100/116

101
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Additional Revenue loss calculation for screenwriters, directors & other co-authors
Source: PMP Strategy analysis, FERA, SAA, European Audiovisual Observatory, Experts’ interviews
‘23 AV production market size
(budget) and ‘28 forecasts
Gen AI ’28
cannib. Rates by segments
’28 creators’ revenue
loss (and %)
Share of main segments going
to AV creators x x =
€209Bn €229Bn
’23 market size ’28 forecasts
Includes streamers & broadcasters original
content spending, public fundings and fiscal
incitation (i.e., includes also independent
production)
For directors & other co-authors
c.4-5%
For screenwriters
Includes upfront payment and
contingent payment from producers
c.4-5%
For directors & other co-authors
For screenwriters
• Unscripted works
• Higher production
budget
• Lower production
budget
• Animated works
• Higher production
budget
• Lower production
budget
• Animated works
€1.9Bn
€1.2Bn
B
Revenue
loss 2
15%
20%
Appendix – Detailed methodology and assumptions Audiovisual

101/116

102
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Additional Revenue loss calculation for authors of audiovisual translations,
adaptations and subtitles
Source: PMP Strategy analysis
Gen AI ’28
cannib. Rates by segments ’28 creators’ revenue loss (and %)
x =
’23 worldwide (non CMO-collected)
revenues and ‘28 forecasts
€1.3Bn • Higher production budget €1.9Bn €2.2Bn • Lower production budget
’23 market size ’28 forecasts(1)
Revenue
loss 2
56%
Appendix – Detailed methodology and assumptions Audiovisual

102/116

103
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual Gen AI providers’ revenues – 2023-2028, €Bn
Calculation methodology for the audiovisual segments (1/2) Complete
video/audiovisual works
Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights
Runway 2028 active users'
estimation Runway 2028 ARPU Runway market share in prompt-to-video market
c. €3.5Bn
2028 associated players’
total revenues
c.4m - 30% 40€ 15%
x / =
Video/Audiovisual software market size
2023 2028
€1.4Bn €1.9Bn
2028 Generative AI tools penetration rate
€0.8Bn
2028 associated players
= total revenues
40%(1) / Mid
x
Note: (1) Considered as slightly lower as in audiovisual, due weaker maturity of tools to date, and an estimated even more expensive
price for these tools in 2028
1. Mass public
tools
(also suitable/used
for professional
purposes)
2. Gen AI
tools/software as
an assistant
(support in the
creative process)
Gen AI services 3 revenues
Appendix – Detailed methodology and assumptions Audiovisual

103/116

104
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Audiovisual Gen AI providers’ revenues – 2023-2028, €Bn
Calculation methodology for the audiovisual segments (2/2)
Dubbing/Subtitling and Screenwriting
Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights
Screenwriting software market size
2023 2028
2028 Generative AI tools penetration rate 2028 associated players
= total revenues
€140m €230m 60% / High €125m
x
Dubbing/Subtitling software market size (incl. social media) 2028 Generative AI tools penetration rate(1) 2028 associated players
= total revenues x
€1.3Bn €1.9Bn 41% €0.9Bn
2023 2028
Dubbing /
Subtitling tools
Screenwriting
tools
Gen AI services 3 revenues
Note: (1) Based on weighed penetration rate calculated in the market segment | (2) Higher rate as in the market calculation, as such
software are widely used for audiovisual content that is not intended to generate revenue or is not widely distributed (corporate
content, content for social networks)
Appendix – Detailed methodology and assumptions Audiovisual

104/116

105
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Detailed methodology and assumptions
Glossary
Detailed list of interviews conducted
PMP Strategy presentation
Appendix

105/116

106
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Glossary | Main abbreviations and definitions (1/2)
• AV: Audiovisual
• ARPU: Average Revenue Per User
• B2B (Business-to-Business): Refers to transactions between businesses, such as a manufacturer selling to a wholesaler. Examples include companies providing
office supplies to other businesses
• B2C (Business-to-Consumer): Businesses selling products or services directly to individual consumers. Examples include online retailers like Amazon.
• Buy-out: A one-time payment for the full rights to use a creative work, with no future royalties owed to the creator.
• CAGR: Compound Annual Growth Rate
• CMO: Collective Management Organization
• Deep Learning: A subset of machine learning involving neural networks with many layers, enabling the analysis and learning from large amounts of complex data.
• DSPs (Digital Service Providers): In the context of music streaming, Digital Service Providers are online platforms that distribute and stream music to listeners.
Examples include Spotify, Apple Music, and Amazon Music.
• GAFAM: Acronym for Google, Apple, Facebook, Amazon, and Microsoft
• Gen AI (Generative Artificial Intelligence): AI systems that generate new content based on training data.
• Input: The data or information fed into an AI system or algorithm for processing and analysis.
• LLM (Large Language Model): A type of artificial intelligence model trained on vast amounts of text data to understand and generate human language.
• Machine Learning: A branch of artificial intelligence where algorithms learn from and make predictions or decisions based on data.
• NLP (Natural Language Processing): A field of artificial intelligence focused on the interaction between computers and humans through natural language.
• OTT: Over the top, self-distribution model outside the operator set-top box: content accessible directly through an app/website on all devices (smart TVs,
smartphones, tablets, etc.)
Appendix - Glossary

106/116

107
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
• Output: The result or product generated by an AI system, such as text, images, or other data.
• Pay-per-view (TVoD): a television service in which viewers are required to pay a fee in order to watch a specific programme.
• UGC (User-Generated Content): Content created and published by users rather than by professional creators or brands, often shared on social media and other
online platforms.
• VOD: Video-on-demand
– AVOD: Advertising video on demand, i.e. advertised-funded digital video platforms (YouTube, Social Media)
– BVOD: Broadcaster Video On Demand, free-access streaming platforms from local broadcasters (VRT MAX, VTM Go, Go Play)
– HVOD: Hybrid video on demand, combining several business models (advertised-funded and subscription/consumer-funded for instance)
– SVOD: Subscription video on demand, traditional streaming platforms, including international players (Netflix, Disney+, etc.) and local players (Streamz)
– FAST: Free ad-supported streaming TV
Glossary | Main abbreviations and definitions (2/2)
Appendix - Glossary

107/116

108
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Detailed methodology and assumptions
Glossary
Detailed list of interviews conducted
PMP Strategy presentation
Appendix

108/116

109
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Interviews | CMOs (1/2)
Source: Experts’ interviews & workshops
Name Company Repertoire
Marie-Anne Ferry-Fall & Thierry Maillard ADAGP Visual Arts
Dean Ormston & Richard Mallett APRA AMCO Music
Christian Zimmermann & Reema Selhi DACS Visual Arts
Ricardo Gómez Cabaleiro DAMA Audiovisual
Tobias Holzmüller & Kai Welp GEMA Music
Kazumasa Izawa & Kay Yamaguchi JASRAC Music
Chu Ga Yeoul & Seon Cheol Hwang KOMCA Music
Andrea Czapary Martin & John Mottram PRS Music
Alexandra Cardona Restrepo REDES Audiovisual
Géraldine Loulergue-Husson & Patrick Raude & Sandrine Sandoval SACD Audiovisual
Héloïse Fontanel Sacem Music
Julien Dumon Sacem Music
David El Sayegh Sacem Music
Julien Lefebvre Sacem Music
Appendix – Interviews

109/116

110
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries Source: Experts’ interviews & workshops
Name Company Repertoire
Annabell Lebethe SAMRO Music
Cristina Perpiñá-Robert Navarro SGAE Transversal
Matteo Fedeli & Fabrizio Zavagli & Adriana Galli & Andrea Marzulli SIAE Transversal
Jennifer Brown SOCAN Music
Jürg Ruchti SSA Audiovisual
Marcelo Bastos Castello Branco & collaborators UBC Music
Sylwia Biadun ZAPA Audiovisual
Vianney Beaudeu & Raphaël Léaupard LaScam Audiovisual
Maria Garateche Argentores Audiovisual
Richard Combes ALCS Audiovisual
Interviews | CMOs (2/2)
Appendix – Interviews

110/116

111
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Interviews | Tech players
Appendix – Interviews
Source: Experts’ interviews
Name Company Repertoire
Àlex Loscos BMAT Music
Ed Newton Rex Fairly Trained Transversal
Ryan Groves Infinite Album Music / Audiovisual
Nathalie Birocheau Ircam Amplify Music
Alexandre Défossez Kyutai Transversal
Philippe Guillaud Matchtune Music
Eric Samson Microsoft Transversal
Christophe Müller & Kevin Montler YouTube / Google Transversal
Appendix – Interviews

111/116

112
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Interviews | Production, Distribution, Publishing companies
Source: Experts’ interviews
Name Company Organisation type
Pierre-Michel Levallois BAM Music Production Company
Aurélien Hérault Deezer DSP
Mathieu Taieb Dubbing Brothers Production Company
Perrine Guyomard Ex-Warner / Sacem Lab Production Company / CMO
Tiphaine Des Déserts Getty Image Production & Publishing
Michael Turbot Sony Computer Science Laboratories Production Company
Anne Jouanneau Sony Music Publishing Publishing Company
Appendix – Interviews

112/116

113
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Interviews | Institutions, Legal bodies and other organisations
Source: Experts’ interviews
Name Company Repertoire
Marion Carré Ask Mona / Commission Européenne Transversal
Sylvie Fodor CEPIC Visual Arts
Cécile Lacoue CNC Audiovisual
Arshia Cont Ex-Ircam / Antescofo Music
John Phelan ICMP Music
Lauri Rechard, Abbas Lightwalla IFPI Music
Alfons Karabuda NIM / ECSA Music
Alexandra Bensamoun Paris Saclay / Commission interministérielle de l'IA Transversal
Benoît Carré SGYGGE / Ministère de la Culture Music
Isabelle Wekstein-Steg WAN AVOCATS Transversal
Juliette Prissard Eurocinema Audiovisual
Céline Despringre SAA Audiovisual
Pauline Durand-Vialle FERA Audiovisual
Gilles Fontaine European Audiovisual Observatory Audiovisual
Eduardo Senna & Matheus Leopardi Senna Advogados Audiovisual
Appendix – Interviews

113/116

114
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Detailed methodology and assumptions
Glossary
Detailed list of interviews conducted
PMP Strategy presentation
Appendix

114/116

115
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Your business environment is changing faster and faster. We already
know that we will not work or consume tomorrow as we do today.
We have learned that we need to be able to adapt quickly to major
disruptions, which are unpredictable by nature. And that it is no longer
acceptable to impact business performance without taking into
consideration the world we live in.
At PMP Strategy, we are committed to working with senior executives
to achieve both goals and create a positive impact.
Just as you are, we are true experts in your market. We apply rigorous
analysis, leveraging our high level of competence and understanding.
We believe that there is no positive impact without great conviction
and the total mobilisation of a diversified team.
We cannot address today's problems with yesterday's answers. We are
dedicated to finding tailor-made and innovative solutions, as if we
were doing it for ourselves. We are always focused on working with
you and your teams, hand in hand, with the entrepreneurial spirit
that motivates us.
We strongly believe in diversity, human commitment and openness.
These are the values that forge the strong relationships and mutual
trust that we cultivate with our clients.
PMP Strategy is a renowned consulting firm in the Media & Cultural
industries, helping its clients adapt their business models and
organisations in a market disrupted by digital "pure players".
A Strategic Consulting Firm
Telecom Media Tech
Transportation & Mobility
Energy & Industry Decarbonization
Financial Services & Institutions
Strategy & Transformation
Private Equity
Digital, Data & Customer Experience
CFO Advisory and Integrated Performance
CSR-ESG
Industries Capabilities
PMP Strategy: An International Player
Europe Africa & Middle East North America
Dubai, Casablanca New York, Montreal,
Seattle, Toronto
Key figures
11 Offices. +150 Consultants.
+100 Expert Advisors.
Paris, London,
Madrid, Brussels,
Luxembourg
Cultural & Creative Industries

115/116

116
Study on the economic impact
of Generative AI in the Music and
Audiovisual industries
Philippe Curt
pcurt@pmpstrategy.com
Helene Moin
hmoin@pmpstrategy.com
François Cousi
fcousi@pmpstrategy.com
Raphaël Flabeau
rflabeau@pmpstrategy.com
CISAC Communications
+33 1 55 62 08 50
communications@cisac.org
Contacts

116/116

Study on the economic impact of Generative AI in the Music and Audiovisual industries

1. 1 Study on the economic impact of Generative AI in the Music and Confidential PMP Strategy Audiovisual industries Study on the economic impact of Generative AI in the Music and Audiovisual industries Complete study Current situation and 5-year perspective November 2024 Confidential PMP Strategy

2. 2 Study on the economic impact of Generative AI in the Music and Audiovisual industries 1 Introduction: Study context & objectives p.4 2 Generative AI overview p.10 What is Generative AI? p.11 Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? p.20 What are the issues at stake in terms of copyright management? p. 24 What are the main trends driving the growth of Generative AI in creation, today and by 2028? p.31 3 Economic impact in Music and Audiovisual creation p.39 What will be the economic impact of Generative AI in Music by 2028? – use cases and economic estimates p.51 What will be the economic impact of Generative AI in Audiovisual by 2028? – use cases and economic estimates p.72 Approach and methodology p.40 Content of the study

3. 3 Study on the economic impact of Generative AI in the Music and Audiovisual industries Disclaimer 1 This Study was prepared by PMP Strategy, an independent strategy consulting firm mandated by CISAC, to assess the economic impact of Generative AI on creation in the Music and Audiovisual sectors. 2 The study provides PMP Strategy’s independent and objective view on the evolution and impact of the use of Generative AI services on the two repertoires considered up to 2028. The historical figures and forecast assumptions are based on market data, relevant benchmark and interviews with industry experts: Collective Management organizations (CMOs), creators, tech players, producers, publishers, DSPs, and institutional players representative of the two industries. 3 Inevitably, unanticipated events and circumstances may occur, and some of the assumptions used to develop the forecasts may not be realized. Consequently, while we consider that the information and opinion given in this Report are sound, PMP Strategy does not guarantee or warrant the conclusions contained in the Report. 4 The Study is valid at the date of completion, which may fall prior to publication. The authors do not take responsibility for any information or events after the Report’s delivery date which may affect its contents.

4. 4 Study on the economic impact of Generative AI in the Music and Audiovisual industries 1. Introduction Context, objectives, methodology

5. 5 Study on the economic impact of Generative AI in the Music and Audiovisual industries CISAC has commissioned PMP Strategy to assess the economic impact of Generative AI on creation in the Music and Audiovisual sectors Market size 1 What will be the market size of Music and Audiovisual outputs generated by AI in 5 years (2028)? Revenue loss 2 Potential cannibalisation of creator’s revenue streams due to the substitution of human works by Gen AI outputs What will be the associated loss of revenue for creators by 2028? Market penetration and market value (on both B2C and B2B segments) of Gen AI outputs Gen AI services’ revenues 3 Revenues of Gen AI tools aimed at the general public and professionals, offering either complete outputs generation and/or assistance in the creative process What will be the revenues of Gen AI tools/services providers by 2028? Introduction: Context, objectives, methodology

6. 6 Study on the economic impact of Generative AI in the Music and Audiovisual industries The evaluation focuses on (1)the value of Gen AI outputs in the market, 2) The evaluation focuses on(1) the associated impact on creators' revenues, The evaluation focuses on (1) the revenues of the tools enabling outputs’ generation Source: PMP Strategy analysis 1 2 3 CREATION DISTRIBUTION/BROADCASTING CONSUMPTION Gen AI tool/service Gen AI service user Distributor / Broadcaster Subscription/pay per act Subscription / pay per act / advertising Creators Creators’ revenue loss 2 End-user Collections CMOs Other revenue streams Introduction: Context, objectives, methodology Legend Gen AI output flow Money flow Gen AI tools revenues’ 3 Gen AI Outputs’ market size 1

7. 7 Study on the economic impact of Generative AI in the Music and Audiovisual industries Two-level impacts analysis Impact of market penetration by Gen AI outputs and impact on creators' revenues (both in terms of revenue loss due to cannibalisation and of revenue opportunity) Broad scope Broad scope of the study: 2 repertoires, international footprint Involvement of industry experts Strong involvement of CMOs and representatives from the industries: 50 industry professionals interviewed or involved in workshops Use cases & market trends analysis Detailed analysis of use cases and underlying factors / market trends determining their evolution over the next 5 years Transparent methodology Transparent methodology and assumptions built and validated with CMOs and CISAC team Quantitative and qualitative analysis, based on interviews and insights from industry experts, existing studies / market data and workshops Exhaustive approach The study aims at identifying the main applications of Gen AI in these fields by 2028 and estimating their economic impact Introduction: Context, objectives, methodology

8. 8 Study on the economic impact of Generative AI in the Music and Audiovisual industries Introduction: Context, objectives, methodology The study is based on experts' interviews, internal and external data analyses, and workshops …interviews with industry professionals from the Music, Visual Arts and Audiovisual sectors between July and September 2024 +50 (Creators, Producers, Publishers, Distributors, DSPs, CMOs, Tech & AI companies, institutional players) Expert interviews Public and private players’ data sources Data sources • Market data • Studies and panels on use cases and trends in Generative AI • Literature and main texts on regulatory context and copyrights issues Workshop sessions with CISAC members and industry experts 8 Workshops

9. 9 Study on the economic impact of Generative AI in the Music and Audiovisual industries 50 interviews were conducted with representative stakeholders from the 2 industries, across the value chain Introduction: Context, objectives, methodology Music Audiovisual Other institutions

10. 10 Study on the economic impact of Generative AI in the Music and Audiovisual industries 2. Generative AI overview

11. 11 Study on the economic impact of Generative AI in the Music and Audiovisual industries Generative AI overview What is Generative AI? Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? What are the issues at stake in terms of copyright management? What are the main trends driving the growth of Generative AI in creation, today and by 2028? Generative AI overview

12. 12 Study on the economic impact of Generative AI in the Music and Audiovisual industries Artificial Intelligence (AI) Machine learning supervised or unsupervised Deep learning (1)Large language models, enabling to perform a variety of natural language processing tasks (generate human-like text, classify text, answer questions, etc.) Associated functionalities What are the main characteristics & specificities of each technology? Ex. of applications Ability given to machines to mimic human intelligence and cognitive functions to perform various tasks (problem solving, learning) n.a. (research field) How do they translate concretely? Ability given to machines to learn from structured datasets, without explicit programming, to detect patterns and make decisions/predictions Customer behaviour prediction Weather forecasting Ability given to machines to learn complex patterns based on artificial neural networks, from large and unstructured datasets Facial recognition Documents reading Ability given to machines, notably relying on LLMs(1), to learn complex patterns to generate new content: language, visual and audio, etc. Image/video creation Human-like interaction AI fields What are the different AI technologies? How do they interlock with each other? 1950’s 1970’s 1980’s 2000’s 2010’s 2020’s Key events popularizing AI technological advances 1996 G. Kasparov, world’s chess champion, was defeated by Deep Blue, IBM’s developed supercomputer relying on machine learning 2011 Watson, IBM’s developed supercomputer relying on deep learning, won 1st place in US quiz show Jeopardy! against historic champions 2022 OpenAI publicly launches ChatGPT, an AI-powered chatbot engaging in conversational dialogues and providing responses to user queries Generative AI Source: Specialized press, Stanford University, PMPS Analysis Generative AI is the recent pinnacle of 50 years of progress in Artificial Intelligence What is Generative AI?

13. 13 Study on the economic impact of Generative AI in the Music and Audiovisual industries 3.1. Prompt • Instruction of a prompt by the user to the model (in the form of text, image, video or audio) 3.2. Prompt processing • Processing of the prompt by the AI program, leveraging the training phase 3.3. Output generation • Generation of an output (new content) in the form of text, visual or audio content AI model production and use Key steps for the development and use of a Generative AI model Generative AI models leverage deep learning on large datasets to generate new content (image, text, video, audio) upon the user’s instruction Source: Specialized press, PMPS Analysis 1.1. Data collection: • Selection of large datasets, relevant to the type of output to be generated by the program • Necessary mass copying and storage of data 1.2. Data preprocessing: • Preparation of the raw data for analysis (cleaning, normalising, labelling, enhancing, etc.) 2.1. Model architecture building • Selection and building of the model architecture (including GANs, VAEs, transformer-based : see detail on next page) 2.2. Model training • Training of the model, taught from the pre-processed dataset • Unsupervised learning 2.3. Model optimization • Continuous/iterative performance evaluation • Adjustments/refinement of parameters (to minimise difference between the output and real data) • Improvement of the model structure AI model development 1 – Collection & preprocessing 2 – Training 1.1. Data (input) collection & 1.2. pre-processing 2.3 - Model optimization 2.2 - Model training 2.1 - Model architecture building 3 – Content generation 3.1 - Prompt 3.2 – Prompt processing 3.3 - Output generation What is Generative AI?

14. 14 Study on the economic impact of Generative AI in the Music and Audiovisual industries Most Gen AI programs today are based on 3 models - GANs, VAEs and transformer-based models, with specific applications and benefits Source: Specialized press, PMPS Analysis • DeepArt.io: Transforms user-uploaded photos into artwork in the style of famous painters using GANs • ThisPersonDoesNotExist.com: Generates lifelike human faces that don't belong to any real individuals Involvement of two neural networks in GANs: • The generator, which creates data (produces data so convincing that the discriminator cannot distinguish it from real data) • The discriminator, which evaluates it (becoming better at identifying fake data over time) • Image creation • Realistic photographs generation • Art, and fashion designs • Video game environments • … Generative Adversarial Networks (GANs) Description Main applications Examples of AI Services The field is rapidly evolving, with new models being developed regularly – other models include autoregressive, diffusion models, RNNs, EBMs, and flow-based models • Jukebox by OpenAI: Produces music in various genres and styles by sampling and processing audio in latent space • AIVA (Artificial Intelligence Virtual Artist): Composes original music scores suitable for films, games, and other content Two key phases in the VAEs’ generative model: 1. Encode input data into a latent space 2. Decode to generate new, similar data Learning of complex data distributions and producing new instances similar to the input data • Image generation • Synthetic datasets creation • Drug discovery • Music generation or other audio content • … Variational Autoencoders (VAEs) Use of attention mechanisms to process sequences of data (text or pixels), by focusing on different parts of the data at different times Generation of coherent and contextually relevant content • Natural language tasks (translation, summarization, and text generation) • Image-related tasks • … Transformerbased Models • GPT-3 by OpenAI: An advanced language model capable of understanding and generating human-like text (answers to questions and creates content) • DALL-E by OpenAI: Generates imaginative images and art from textual descriptions What is Generative AI?

15. 15 Study on the economic impact of Generative AI in the Music and Audiovisual industries Text-based prompt Audio + Lyrics Generation process of Gen AI outputs Generative AI engines can handle all data formats to generate increasingly diverse contents, and perform a wide range of tasks Source: Specialized press, Suno website, PMPS Analysis 3.1 - Prompt 3.3 - Output generation Example: 3.2 – Prompt processing 3.1 - Prompt 3.3 - Output generation 3 – Content generation 3.2 – Prompt processing 3.1 - Prompt 3.3 - Output generation Question answering Sentiment Analysis Information extraction Image Captioning Object Recognition Instruction Following Images-based prompts Audio-based prompts Text-based prompts Mixed-media prompts Text output Image output Audio output Mixed-media output Video output > Video-based prompts Others What is Generative AI? Prompt format Task Output format

16. 16 Study on the economic impact of Generative AI in the Music and Audiovisual industries Illustration of Midjourney technology performance evolution | 2022 – 2023 Recent cases have demonstrated the ability to generate content always closer to human creations Source: All Midjourney Versions (V1-V6) Compared: The Evolution of Midjourney - Aituts Midjourney V1 Midjourney V2 Midjourney V3 Midjourney V4 Midjourney V5.1 Midjourney V6 Initial version with raw results Feb 2022 April 2022 July 2022 Nov 2022 Mai 2023 Introduction of upscaling and variations, improved coherence Improved lighting, reflections and realism. Added stylised and quality parameters Photorealistic quality, ability to generate complex designs V5.1 to V5.2: Greater realism and aesthetics Text & Image-based prompt Image output vintage photo, girl smoking cigarette, irina nordsol kuzmina, a hazy memory, pixiv --ar 2:3 Image Text prompt Dec 2023 Improved image quality and prompt understanding What is Generative AI?

17. 17 Study on the economic impact of Generative AI in the Music and Audiovisual industries Extract of Qianqiu Shisong, China Media Group AI-generated animated series | 2022 Generative AI tools are thus increasingly questioning the very notion of creation Source: Experts’ interviews, specialized press • In February 2024, Chinese state broadcaster, China Media Group (CMG), launched the country’s first animated series created with a Generative AI tool, Qianqiu Shisong, which features ancient stories based on traditional Chinese poems and verses, and aims to showcase the country’s traditional culture and aesthetics. • The series was produced using CMG’s internal text-to-video model (Media GPT), trained on traditional Chinese poetry and video and audio material from China Media’s catalogue. • The production studio indicated that artificial intelligence was used at every step of the development and production process, from design to video generation and post-production. What is Generative AI?

18. 18 Study on the economic impact of Generative AI in the Music and Audiovisual industries 0.8 2.1 4.2 2.9 25.2 2019 2020 2021 2022 2023 22.5 0.7 0.6 1.4 2023 25,2 Private investment in Generative AI | Worldwide, 2029 – 2023, $Bn While total investment in AI has recently slowed down, the Generative AI market is skyrocketing, with an unprecedented surge in 2023 highly driven by the US Source: Stanford University – AI index report 2024, PMPS Analysis x9 Other China EU + UK USA Worldwide • Investments in Gen AI have surged recently as the technology demonstrates its potential to transform industries and reshape the business landscape • A wide range of startups and Gen AI applications are targeted by investments in sectors such as technology, telecom, healthcare, financial services, energy, consumer goods, media, culture, and entertainment • Generative AI is becoming a key driver of innovation, with applications that enhance operational processes and create new products and services, impacting nearly every aspect of the modern economy Over the past few years, we've witnessed a significant surge in investments in Generative AI by major tech companies and private investors. This trend is driven by the potential of Generative AI to revolutionize jobs in various sectors. Tech Company > What is Generative AI?

19. 19 Study on the economic impact of Generative AI in the Music and Audiovisual industries Generative AI encompasses a fragmented and fast-developing ecosystem, with major generalist players mostly related to GAFAM and multiple smaller solutions serving specific purposes Source: Specialized press, PMPS Analysis The ecosystem has seen an exponential growth in the last year, and is polarized around a few mature and powerful big players, mainly related to GAFAM (~1bn visitors/month on OpenAI.com), and a very scattered network of small and specialized newcomers Generative AI services mapping (non-exhaustive) Specialists Generalists Mature GPT-3.5 Microsoft (Bard) - Google Microsoft Newcomers Amazon / Google Elon Musk Meta Salesforce Google What is Generative AI? 19 Study on the economic impact of Generative AI in the Music and Audiovisual industries

20. 20 Study on the economic impact of Generative AI in the Music and Audiovisual industries What is Generative AI? Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? What are the issues at stake in terms of copyright management? What are the main trends driving the growth of Generative AI in creation, today and by 2028? Generative AI overview

21. 21 Study on the economic impact of Generative AI in the Music and Audiovisual industries Creative inspiration / Research and planning Creation / Execution In Music and Audiovisual, Gen AI services have emerged with use cases ranging from assistance on specific tasks to fully automated complete outputs generation Source: Experts’ interviews, specialized press Post-production (1) Mainly for general users (entertainment) | (2) Mainly for artists/professionals | (3) Not available yet Audiovisual Music Full execution / composition(1) Support to creation(2) iMyFone VoxBox (3) Gen AI ecosystem in creative industries

22. 22 Study on the economic impact of Generative AI in the Music and Audiovisual industries The ecosystem in these fields is mainly made up of very recent, fast-growing newcomers(1) Source: Specialized press, PMPS Analysis Note: (1) As of July 2024 2016 2017 2018 2019 2020 2021 2022 2023 2024 Sora Make-A-Video by AudioCraft by • New ‘Suno for Mobile’ launch in July 2024 • 12 million total users (1) • Approx.. 860k tracks generated/day (1) Gen AI players (non-exhaustive) Gen AI ecosystem in creative industries

23. 23 Study on the economic impact of Generative AI in the Music and Audiovisual industries Models vary according to the use cases addressed and target audiences, and have not all yet reached their full level of maturity Source: Experts’ interviews, specialized press Consumer Prosumer Professional Main Players Complete prompt-to-output tools Assistance in the creative process Destinated to mass public use Destinated to industry professionals Note: (1) Experienced amateurs • Mainly newcomers, tech startups • Newcomers / tech startups • New tools / functionalities of traditional players Services / Key features • Mainly complete prompt-to-output generation (images, video, music) • Gen AI tools / functionalities to automate specific tasks :, mastering, editing, image enhancement, … • Prompt-to-output generation • Gen AI tools / functionalities to automate specific tasks Economic model/ offer • Mainly free-use and freemium (subscription-based) models • Revenues driven by ads, external investments and subscriptions (freemium models) • Mainly Licenses (packages’ purchase) models • Free-use and freemium (subscription-based) models • Licenses models StyleGAN Consumer tools, also suitable for professional use Prosumer tools Professional tools/software Gen AI ecosystem in creative industries

24. 24 Study on the economic impact of Generative AI in the Music and Audiovisual industries What is Generative AI? Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? What are the issues at stake in terms of copyright management? What are the main trends driving the growth of Generative AI in creation, today and by 2028? Generative AI overview

25. 25 Study on the economic impact of Generative AI in the Music and Audiovisual industries The use of Gen AI in creative industries raises two main issues related to copyright management AI models’ inputs AI models’ outputs • To what extent AI models have been trained by using copyrighted works in their datasets? • How can creators be remunerated for the use of their works to train Gen AI models? • What will be the implications of the increasing use of synthetic data? • Do Gen AI outputs infringe copyright on existing works? • e.g., creation of works “in the style of” • Who is/should be liable in case of copyright infringement? • What could be the ownership and “copyrightability” of Gen AI content? • What can/should be considered as Gen AI content? • Should Gen AI outputs be protected? Who would own the rights on Gen AI outputs? Source: Experts’ interviews, specialized press, PMPS Analysis Copyright issues & current regulatory framework

26. 26 Study on the economic impact of Generative AI in the Music and Audiovisual industries SUNO vs. JEN AI vs. Genario: a comparison of copyrights concerns along the Gen AI supply chain There are widely varying degrees of transparency in the operation and construction of AI models Specialized press, PMPS Analysis When asked about OpenAI’s training data sets, CTO Mira Murati responded: “We used publicly available data and licensed data”. However, publicly available data doesn’t mean copyright-free data (ex: Youtube videos). Source: • No transparency on dataset • No compensation / licensing of the data used for training • No publication or declarations on model functioning • Outputs’ ownership to Suno for free subs., and to users for Pro/Premier subs • No verification of the outputs • Model trained on licensed catalogues (Fairly Trained certif.) • No licenses with creators’ representatives • Publication of research paper on how the model works • Outputs’ ownership to the user • Audio recognition and copyright identification of outputs 3 – Content generation AI model Use 3.2 – prompt processing 3.1 - Prompt 3.3 - Output generation 1 – Data collection & preprocessing 2 – Model training AI Model Development 1. Data (input) collection & pre-processing 2.3 - Model optimization 2.2 - Model training 2.1 - Model architecture building • License with SACD (audiovisual CMO) to remunerate authors for the use of their works in the models’ training • Publication of research paper on how the model works • Outputs’ exclusive ownership to the user • No outputs used to train the model Copyright issues & current regulatory framework

27. 27 Study on the economic impact of Generative AI in the Music and Audiovisual industries The performance of Generative AI models is closely tied to their training datasets, ranging from in-house data to AI-generated synthetic data, raising copyright concerns Source: Experts’ Interviews, Specialized press, PMPS Analysis Catalogues & Players (non-exhaustive) Internal user generated content/data In-house data User-generated content Used to tailor models to specific user base External data source Public data/Open source Used for the quantity of resources available Public datasets Open-source repositories Available via Proprietary Used for the quality of the catalogue Copyrighted works catalogues AI-generated data source Synthetic data source AI-generated content Used for the constrained-free offer : logistics, privacy and copyright etc. • As synthetic datasets become more widespread, transparency issues in the training process of Gen AI models are becoming increasingly urgent • However, AI models will always need non-synthetic, human-made data for bias mitigation, renewed creativity and staying in touch with current trends Specificities & risks • Highly relevant and personalized datasets • Enables to improve user engagement and satisfaction • Privacy concerns • Data bias reflecting user base • Broad coverage and cost effective • Promotes transparency and reproducibility • Variable quality and reliability • Potential for outdated or irrelevant data • High accuracy and reliability • Domain-specific insights • Expensive data • Licensing restrictions • Potential ethical concerns • Cost and time efficient (generated faster and more affordably) • Non-copyrighted material • Data availability • Consistency and control Copyright issues & current regulatory framework

28. 28 Study on the economic impact of Generative AI in the Music and Audiovisual industries Industry players have started taking legal action against Gen AI services for the unauthorized use of their catalogues in models training Source: Specialized press, Music Business Worldwide, Experts’ interviews, PMPS Analysis • In June 2024, Universal, Warner and Sony filed federal copyright infringement lawsuits against AI music generator platforms Suno and Udio • The Majors accuse the platforms of "mass infringement of copyrighted sound recordings copied and exploited without permission.“ Major music groups vs. Suno & Udio VS. 1. • The New York Times filed a lawsuit against OpenAI and Microsoft for copyright infringement, claiming that millions of its articles were used without authorization • These works were allegedly used to train AI technologies like ChatGPT, which now compete with the newspaper as a source of reliable information New York Times vs. ChatGPT VS. 3. • Universal Music, Concord and Abkco have filed a copyright infringement lawsuit against the AI start-up Anthropic • They accuse the AI company of using their artists' lyrics without permission to generate near-identical copies through its AI model Claude Majors & music groups vs. Anthropic 2. VS. • A trio of artists (Sarah Andersen, Kelly McKernan and Karla Ortiz) filed a lawsuit against AI companies Stability AI and Midjourney for copyright infringement • Several artists have joined the legal action, including other AI services in the prosecution: DeviantArt and Runway AI 4. Artists vs. Stability & Midjourney VS. Other examples Music • In addition to legal action taken against Gen AI services, CMOs have started to establish opt-out mechanisms to prevent the training of Generative AI models using copyrighted works of their members (e.g., Sacem for Music) • However, these mechanisms only apply to future training of AI models and are made possible by laws mandating transparency in the training processes of AI models Copyright issues & current regulatory framework

29. 29 Study on the economic impact of Generative AI in the Music and Audiovisual industries United States Europe Rest of the world Overall, the regulatory framework is still in progress and remains heterogeneous across regions Source: Expert interviews, specialized press Generative AI Copyright Disclosure Act (April 2024): A proposed law requiring AI companies to submit a list of all copyrighted works used for training their AI models. US Copyright Office Guidelines (March 16, 2023): Clarified the necessity of human contribution to qualify for copyright protection, stressing that tools like AI can be part of the creative process, but human control over the expression is essential. Directive on copyright and neighbouring rights in the digital market (April 2019): Directive allowing text and data mining (TDM) necessary for AI training under certain conditions: • Article 3: Allows data mining for scientific purposes without special conditions • Article 4: Allows data mining for all purposes, including commercial, provided access to the data was lawful and rightsholders did not opt out AI Act (April 16, 2024): Introduces several obligations for AI systems: • Ensuring respect of copyright, including for opensource foundational models • Publishing detailed summaries of works used for AI training • Identifying AI-generated content as such • Extraterritorial application, effective from August 1, 2024, with phased implementation until 2027 Council of Europe Framework Convention on AI (May 17, 2024): Focuses on respecting human rights in AI development, emphasizing transparency for enforcing intellectual property rights. Guidance for Gen AI in education and research (UNESCO, Sept. 2023): Calls for immediate actions and long-term policies to regulate the use of Gen AI in education & research, including text, image, video and music generation. Copyright Act – Art. 30-4 (Japan, May 2018): Copyrighted works can be used in the training of AI models without requiring a license. Rightsholders do not have the option to opt out, and there is no obligation for transparency Gen AI governance framework (Singapore, March 2024): Advises policymakers to clarify the application of existing personal data laws to Generative AI. Aims to foster trusted Generative AI development. AI policy on regulations & ethics (Israel , 2023): Focuses on responsible AI innovation. Emphasizes on “soft regulation” with sector-specific guidelines. Aims to respect the rule of law, fundamental rights and public interests. Non-exhaustive list Copyright issues & current regulatory framework Legend Mainly focused on input issues Mainly focused on output issues

30. 30 Study on the economic impact of Generative AI in the Music and Audiovisual industries Tools are also being developed to support industry players by helping them identify copyrighted works used as inputs, and detect Generative AI outputs Source: Expert interviews, Specialized press Example of tools Deezer is currently developing tools to: • Identify whether a Gen AI model has been trained on specific tracks • Detect music tracks generated by the world's largest music-generating LLMs Detection of copyrighted works used as inputs Objective: Analyse Gen AI models to detect if copyrighted content has been used in their training Method: Compel models to provide specific copyrighted works as outputs, with prompts designed to induce hallucinations, thus proving they have been used in the training process Challenges: Difficult to scale/industrialize the process Detection of outputs generated by AI tools Objective: Scan specific works or entire catalogues to identify whether they have been generated by AI Method: Identify Gen AI models biases and patterns (usually specific to each model) and scan images/music/video catalogues to identify whether they have been AI-generated Challenges: Detection tools need to be regularly trained on popular Gen AI models to ensure they remain performant Spawning AI is developing solutions to help identify whether a visual work has been used as Gen AI tools inputs (Have I been Trained?), help block AI web scrapping and enforce opt outs Ircam Amplify has developed a tool (AI-Generated Detector) allowing to identify and tag Gen AI musical outputs Copyright issues & current regulatory framework AI models’ inputs AI models’ outputs

31. 31 Study on the economic impact of Generative AI in the Music and Audiovisual industries What is Generative AI? Who are Gen AI services’ providers in the field of creative industries and how is the ecosystem structured? What are the issues at stake in terms of copyright management? What are the main trends driving the growth of Generative AI in creation, today and by 2028? Generative AI overview

32. 32 Study on the economic impact of Generative AI in the Music and Audiovisual industries The adoption of Generative AI tools and outputs in creative industries will be determined by the strategies and behaviour of players across the value chain Source: interviews with industry professionals, Specialized press, PMP Strategy analyses Main drivers Impact on Gen AI adoption Positioning in the value chain Market drivers and underlying trends Technological progress • Rapid technical evolution of Generative AI models is expected to continue in the coming years, enhancing their capabilities beyond simple text or image generation to more complex, multi-modal outputs • Higher quality, more diverse, and personalized outputs, opening new opportunities across multiple industries Gen AI providers/Tech companies Growth of the creator economy • Continued growth of user-generated content on social media, fostering the adoption of Gen AI tools to support content creators • Further reduction of the barriers to entry for creation in all creative industries driven by Gen AI tools Creators Evolution of consumer habits • Growing demand for interactive, on-demand, and contextually relevant content reshaping consumption • Increasing trend toward passive content consumption, where digital platforms curate and recommend content to users rather than users actively selecting it themselves, with Gen AI likely to play a pivotal role in driving this shift Consumers/ End-users Strategy and positioning of traditional players • Shift in traditional players' strategies and positioning to adopt Generative AI for competitiveness, in all industries • Integration of Generative AI by players across all segments of the creative industries' value chain: to introduce new offers, optimise content production and distribution, and renew business models Distributors/ Broadcasters & B2B players Regulatory environment and ethical issues • Evolving regulatory frameworks addressing intellectual property, data privacy, ethics, and cultural diversity issues as Gen AI becomes more widespread, potentially impacting its growth • Increased awareness of end users regarding ethical issues related to copyright, fair pricing, and the proper remuneration of authors Legal bodies/CMOs

33. 33 Study on the economic impact of Generative AI in the Music and Audiovisual industries Technological progress Test scores of AI systems on various capabilities relative to human performance | 1998 - 2022 Recent technological progress have enabled AI and Gen AI models to outperform human performance in all basic capabilities, laying the foundation for continued progress in the coming years Source: Kiela et al. (2023) & Our World in Data, PMP Strategy analysis Within each domain, the initial performance of the AI is set to –100. Human performance is used as a baseline, set to zero. • AI performances are above human for every basic (non-complex) capabilities analysed • The later a capability started to be implemented, the faster it reached human-level performance: – Speech recognition: 19 years – Handwriting recognition: 17 years – Image recognition: 7 years – Reading comprehension: 1 year – Language recognition: less than 1 year • Technological advancements have paved the way for Generative AI to revolutionize various sectors and domains, including the creative industries The growth of Generative AI is expected to accelerate even further in the coming years, with widespread adoption and advanced technical capabilities anticipated by 2028, driven by substantial investments in Gen AI models and their associated providers Market drivers and underlying trends -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 Human performance Reading comprehension Language recognition Image recognition Speech recognition Handwriting recognition

34. 34 Study on the economic impact of Generative AI in the Music and Audiovisual industries Social media penetration worldwide and creator economy The significant increase in user-generated content on social media will likely drive the high adoption of Gen AI Source: Specialized press, Experts’ interviews, PMPS Analysis 0% 5% 13% 28% 31% 37% 42% 45% 48% 53% 58% 59% 62% 2000 2005 2010 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 3.05 2.49 2.00 2.00 1.56 • The widespread adoption of social media has driven the growth of the creator economy across all disciplines • The creator economy has been boosted mainly by influencer marketing, content monetization, e-commerce and social selling • Approximately 15% of the 4.2Bn users of social media (both individuals and businesses/brands) worldwide are considered part of the creator economy • These creators, and particularly enterprises, are more inclined to use Gen AI tools to produce more personalized content for their users at scale – e.g., Nike and Coca Cola using Gen AI algorithms to create personalized and engaging advertising campaigns Social media penetration rate | %, Worldwide, 2000-2024 Top 5 social media in number of users | in Bn users Number of players in the creator economy in social media & split by type of players 4.2Bn Social media users Creators Use their influence to aggregate and monetize their audience Creator economy = c.15% of the social media users Of which 2/3 are brands/enterprises (vs. individuals) Passion economy users Engage in any activity to monetize skills on digital platforms/social media Market drivers and underlying trends 500m 200m Creator economy’s growth

35. 35 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approximate average number of tracks uploaded to DSPs each day | thousands, per year, 2018 - 2024 New tools, formats, and distribution channels have significantly lowered barriers to entry in music creation in the last decade, a trend which will be further fostered by Gen AI tools Source: Music Business Worldwide, PMPS Analysis • Music streaming platforms have favoured the creator’s economy in the music field, with: – Lower barriers to entry: easier distribution for independent artists without the need of traditional record label – Increased visibility and reach, through algorithms and curated playlists – Monetization opportunities, through the creation of a new revenue stream with the streaming royalties – Access to data & analytics, community & networking, creative freedom… 20 40 50 60 100 125 143 2018 2019 2020 2021 2022 2023 2024 x3 x2.5 Music streaming providers (non-exhaustive) The streaming platforms and other digital players (e.g., music distributors like CD Baby), combined with the advancements in Generative AI tools, are empowering artists to produce, distribute, and monetize their music more easily than ever before Market drivers and underlying trends Creator economy’s growth

36. 36 Study on the economic impact of Generative AI in the Music and Audiovisual industries Passive music consumption (mood playlists) in streaming platforms | Focus on Spotify Growing demand for interactive, on-demand, contextually relevant content, and the increasing trend toward passive consumption is driving the adoption of Gen AI by streaming platforms (1/2) Source: Music Business Worldwide, PMPS Analysis • Music streaming platforms increasingly offer curated playlists tailored to specific moods and activities, enhancing user engagement • These mood playlists provide a seamless listening experience that requires minimal intervention from the user • End-consumers on streaming platforms are increasingly gravitating towards passive music listening, driven by convenience, personalization, and discovery • In the top 100 Spotify playlists in terms of subscribers, 41% are considered as functional/mood playlists (e.g., Morning Coffee), favouring passive listening Rank Playlists’ name # of subs 1 Today's Top Hits 34M 2 Top 50 - Global 17M 3 RapCaviar 15M 4 Viva Latino 14M 5 Rock Classics 11M 6 Baila Reggaeton 10M 7 All Out 2000s 10M 8 Songs to Sing in the Car 10M 9 All Out 80s 10M 10 Beast Mode 10M Rank Playlists’ name # of subs 89 This Is Michael Jackson 3M 90 This Is One Direction 3M 91 Deep House Relax 3M 92 Hype 3M 93 Hot Hits Deutschland 3M 94 Warm Fuzzy Feeling 3M 95 Coffee Table Jazz 2M 96 Power Hour 2M 97 Intense Studying 2M 98 Chill Vibes 2M Legend Mood playlists/passive listening 41% 59% Commercial music Mood music Market drivers and underlying trends Evolution in consumer habits

37. 37 Study on the economic impact of Generative AI in the Music and Audiovisual industries Video streaming recommendation | Example of Netflix Growing demand for interactive, on-demand, contextually relevant content, and the increasing trend toward passive consumption is driving the adoption of Gen AI by streaming platforms (2/2) Source: Netflix reports, Specialized press, PMPS Analysis • Consumers' appetite for recommended and personalised content is a key driver for the use of Generative AI, enabling even more advanced playlists and content with unlimited tailored content, down to the individual level Diagram on Netflix recommendation system : Collaborative filtering vs. content based Collaborative filtering Content based recommendation Similar users 1 2 Liked by both Liked by her Recommended to him Similar content 1 2 Liked by her Recommended to her • Netflix recommendation’s algorithm includes : – Data collection: viewing history, user interactions, and demographic data – Collaborative filtering: user-user and item-item – Machine learning models : identification of pattern in the user preferences – Content analysis : metadata analyses, natural language processing – Real-time personalization • This directly influences user consumption behaviour: increased engagement and enhanced user experience Market drivers and underlying trends Evolution in consumer habits of Netflix viewership is driven by its recommendation engine 75%

38. 38 Study on the economic impact of Generative AI in the Music and Audiovisual industries Examples of current and projected Gen AI use cases across the main segments of the creative industries' value chain Stakeholders across the entire value chain of the creative industries are increasingly adopting Gen AI to optimize content production and distribution and to renew their value proposition Source: Specialized press, Experts’ interviews, PMPS Analysis • Music: Audiovisual production companies using Gen AI to produce background scores in audiovisual content, lowering production costs, mainly in lower budget works • Audiovisual: Brands using Gen AI video outputs on social media, enhancing end-consumer experience with more tailored and personalized content B2B distributors • Music: Streaming platforms integrating Gen AI outputs in mood playlists, to create more tailored content with no copyright • Audiovisual: VOD platforms using Gen AI videos to create trailers, cutting production costs B2C distributors • Music: Press agencies using Gen AI to create background scores, lowering production costs • Audiovisual: Advertising agencies using Gen AI to create personalized video ads, reducing costs and time Artworks/Content aggregators Commissioners of artworks/content • Music: Music libraries using Gen AI to generate large number of new tracks, increasing catalog options and reducing prices • Audiovisual: Stock photo/video agencies using Gen AI to create short videos content, cutting costs and speeding up availability Traditional Market drivers and underlying trends players’ strategies

39. 39 Study on the economic impact of Generative AI in the Music and Audiovisual industries 3. Economic impact in Music and Audiovisual creation

40. 40 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approach and Methodology What will be the economic impact of Generative AI in Music by 2028? What will be the economic impact of Generative AI in Audiovisual by 2028? Gen AI Economic impact

41. 41 Study on the economic impact of Generative AI in the Music and Audiovisual industries Key figure 1 Key figure 2 Key figure 3 Market size 1 What will be the market size of Music and Audiovisual outputs generated by AI in 5 years (2028)? Revenue loss 2 Potential cannibalisation of creator’s revenue streams due to the substitution of human works by Gen AI outputs What will be the associated loss of revenue for creators by 2028? Market penetration and market value (on both B2C and B2B segments) of Gen AI outputs Gen AI services’ revenues 3 Revenues of Gen AI tools aimed at the general public and professionals, offering either complete outputs generation and/or assistance in the creative process What will be the revenues of Gen AI tools/services providers by 2028? Approach and methodology CISAC has commissioned PMP Strategy to assess the economic impact of Generative AI on creation in the Music and Audiovisual sectors

42. 42 Study on the economic impact of Generative AI in the Music and Audiovisual industries Source: PMP Strategy analysis CREATION DISTRIBUTION/BROADCASTING CONSUMPTION Gen AI tool/service Gen AI service user Distributor / Broadcaster Subscription/pay per act Subscription / pay per act / advertising Creators Creators’ revenue loss 2 End-user Collections CMOs Other revenue streams Legend Gen AI output flow Money flow Gen AI tools revenues’ 3 Gen AI Outputs’ market size 1 Approach and methodology The evaluation focuses on (1)the value of Gen AI outputs in the market, 2) The evaluation focuses on(1) the associated impact on creators' revenues, The evaluation focuses on (1) the revenues of the tools enabling outputs’ generation 1 2 3

43. 43 Study on the economic impact of Generative AI in the Music and Audiovisual industries The methodology relied on qualitative and quantitative analyses, fuelled by interviews with industry players and workshops with CMOs Source: PMP Strategy analysis Use cases identification and prioritisation Qualitative approach • Qualitative analysis used to feed the quantitative part (market hypotheses and impact estimates) • Identification of market segments where generative AI has a significant impact • Translation of qualitative assessments into quantitative estimates • Identification of the most significant generative AI use cases for creation in both fields • Prioritisation of these use cases based on their potential adoption level to determine the ones with the most significant economic impact Economic impact estimation Quantitative approach Approach and methodology

44. 44 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music Generative AI use cases examples in Music and Audiovisual (non-exhaustive) Generative AI has numerous applications in the filed of creative industries and can intervene at all stages of the creative process, from ideation to post-production Source: Experts’ interviews, specialized press, PMPS Analysis Creative inspiration Research & Planning Creation/execution Post-production Distribution & diffusion • Melody generation • Lyrics creation • Trend analysis • Sample discovery • Collaborative composition • Arrangement optimization • Prompt-to-music generation through AI • Automated mixing and mastering • Pitch correction • Voice cloning, synthesis • Playlist recommendation/ optimization • Targeted marketing • Screenwriting exploration • Storyboard generation • Scene visualization • Script analysis • Animation generation • Special effects creation • Video editing • Colour correction • Translating/Adapting • Content recommendation • Audience analysis Scope of the study Audiovisual Approach and methodology

45. 45 Study on the economic impact of Generative AI in the Music and Audiovisual industries More generally, use cases fall into 2 categories: fully automated prompt-to-output applications or assistance in the creative process Source: Experts’ interviews, specialized press, PMPS Analysis Fully Gen AI outputs • Fully automated generation of musical or audiovisual outputs via the use of Gen AI services • No human input beyond the prompt (or marginal intervention) • Creation of musical, visual or audiovisual works with the assistance of Gen AI tools, enhancing human work • Significant human involvement in the creative process (“augmented artist”) AI-assisted work creation > • Prompt-to-video tools such as Sora (OpenAI) or InVideo • Actor rejuvenation, video restoration & colouring, sound, digitalisation, etc. (Respeecher) • Prompt-to-script tools (Genario) • Automated dubbing-subtitling (Veed) > • Prompt-to-song tools such as Suno or Udio • Pitch correction, editing, mastering, etc. (AudioShake, iZotope) • In the market size calculation, only fully Gen AI outputs are considered, as their distribution will affect the market by replacing human-created works. • For the creators’ loss calculation, (i) In the music sector, fully AI-generated outputs will cannibalize creators' revenues in specific market segments; (ii) whereas in the audiovisual sector, complete AI outputs and reduced production budgets due to Gen AI tools (e.g., screenwriting, translation) will lead to revenue losses • There is a grey area where semi-automated works may still be considered as human creations. The study does not aim at estimating the Gen AI contribution in human works Grey area Approach and methodology

46. 46 Study on the economic impact of Generative AI in the Music and Audiovisual industries ▪ What will be the Music and Audiovisual market segments impacted by Gen AI use cases in the next 5 years? ▪ What will be the penetration rate and market value of Gen AI outputs in 5 years? ▪ Can we expect an "AI boost" (additional growth) due to Gen AI? ▪ What will be the share of existing players’ (distributors) revenues driven by Gen AI outputs? Gen AI impact estimation methodology: Market size of Gen AI outputs in 2028 What will be the market size of Music and Audiovisual outputs generated by AI in 5 years (2028)? i.e. market penetration and market value of Gen AI outputs 1 Market size Key questions to be answered Calculation methodology Source: PMP Strategy analysis • Segmentation of the Music and Audiovisual distribution markets (both B2C and B2B – including new Gen AI based services and current distributors) • Estimate of 2023 market size for all the distribution segments likely to be impacted by Gen AI outputs • Forecast to 2028 based on historical growth and market trends • Estimate of Gen AI outputs’ penetration rate for each segment in 2028 based on the prioritised use cases ➢1 Market size of Fully Gen AI outputs by 2028 Approach and methodology

47. 47 Study on the economic impact of Generative AI in the Music and Audiovisual industries ▪ What would be the evolution of creators’ revenues in the next 5 years without Gen AI? - based on current remuneration rules and historical trends ▪ What will be the share of this revenue at risk due to the cannibalisation or substitution of human-made works by Gen AI outputs? Gen AI impact estimation methodology: Creators revenue loss due to Gen AI cannibalisation Revenue loss 2 What will be the associated loss of revenue for creators by 2028? i.e. risk of cannibalisation of creators’ traditional revenue streams Key questions to be answered Calculation methodology Source: PMP Strategy analysis CMO-collected revenues (for both repertoires) • Breakdown of CISAC collections in segments and sub-segments • Estimate of 2023 revenues for each sub-segment and forecast to 2028 • Estimate of cannibalisation rates due to Gen AI outputs by sub-segment Other revenues (only for Audiovisual) • Share of the production budget / dubbing-subtitling market going to audiovisual creators/authors • Estimate of cannibalisation rates due to Gen AI outputs by type of author Approach and methodology 2➢ Potential revenue loss for creators by ‘28 compared to a no Gen AI scenario

48. 48 Study on the economic impact of Generative AI in the Music and Audiovisual industries Gen AI impact estimation methodology: Creators revenue loss due to Gen AI cannibalisation Revenue loss 2 What will be the associated loss of revenue for creators by 2028? i.e. risk of cannibalisation of creators’ traditional revenue streams Source: PMP Strategy analysis Creators’ revenue streams considered for Music and Audiovisual repertoires • The perimeter of rights managed by CMOs is very heterogeneous between regions/geographies • Only in a few countries do CMOs account for a large share of creators’ revenues • For the Audiovisual field, the scope has been extended to capture a better proportion of creators/authors’ revenues Audiovisual creators’ revenue split Music creators’ revenue split • The perimeter of rights managed by CMOs is very homogenous between regions/geographies • CMOs collection account for a significant share of creators’ revenues • For the Music field, the scope for the revenue loss calculation is hence the CMO-collected revenues Approach and methodology CMO-collected rights Impact of Gen AI on CMO-collected rights Rights from upfront payments and other revenues Gen AI impact on upfront payments and other revenues Legend

49. 49 Study on the economic impact of Generative AI in the Music and Audiovisual industries ▪ What will be the evolution of the Gen AI ecosystem in the Music and Audiovisual fields by 2028? ▪ What will be the market penetration of AI-assisted music and audiovisual/video creation tools among professionals by 2028? ▪ How many fully Gen AI promptto-outputs tools will exist by 2028, and what will be their user base and pricing? ▪ What will be the overall revenue generated by both AI assistance and fully prompt-to-output tools by 2028? Gen AI impact estimation methodology: Revenues of Gen AI services Gen AI providers’ revenues 3 What will be the revenues of Gen AI tools/services providers by 2028? Key questions to be answered Calculation methodology Estimated revenue of Gen AI tools and services by 2028 Source: PMP Strategy analysis Approach and methodology AI-assisted creation tools • Estimate of the professional Music and Audiovisual software markets (editing, post-production…) in 2023, forecast to 2028 • Gen AI penetration rate on this segment Full prompt-to-outputs tools • Number of services providing fully automated prompt-to-music tools • Forecast of the average number of users and average revenue per user to 2028 3

50. 50 Study on the economic impact of Generative AI in the Music and Audiovisual industries Gen AI impact estimation methodology: Revenues of by Gen AI services Gen AI providers’ revenues 3 What will be the revenues of Gen AI tools/services providers by 2028? Gen AI tools and service providers have been split in 2 categories Source: PMP Strategy analysis • Prompt-to-video complete outputs generator • Prompt-to dubbing Gen AI providers • Prompt-to-scripts Gen AI providers • Gen AI tools for video ideation, mastering, editing, post-production… • Includes Gen AI tools providing prompt-tovideos as one of their services, but mainly for professionals Audiovisual Full prompt-to- outputs tools AI-assisted creation tools • Prompt-to-songs generator • Gen AI tools for music ideation, mastering, editing, post-production… • Includes Gen AI tools providing prompt-to-music as one of their services, but mainly for professionals Music Approach and methodology

51. 51 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approach and Methodology What will be the economic impact of Generative AI in the Music field by 2028? Main Applications What will be the economic impact of Generative AI in Audiovisual by 2028? 2028 forecast Economic impact in Music creation

52. 52 Study on the economic impact of Generative AI in the Music and Audiovisual industries Identification of Gen AI main applications in the music creation process Source: Experts’ interviews, specialized press, PMP Strategy analysis Creative inspiration / Research and planning Creation / Execution Post-production Main use cases AI-assisted or fully AI-generated creation/composition AI-assisted post-production AI-assisted music conceptualization& idea generation AI-assisted idea generation: Complete / partial musical work creation (samples, moods, rhythms, etc.) used as inspiration for human created work AI-assisted Voice Cloning to create original musical works Voice synthesis/cloning Out of scope (linked to performers’ rights mostly) AI-assisted Pitch correction, Editing, Mastering Automated complete musical output generation for core content in audiovisual works (e.g., video games main theme) Automated complete musical output generation for background content (in AV works, in-store sonorization, social media…) Automated complete musical output generation for commercial distribution on music streaming platforms 5 6 7 Automated complete musical output generation and consumption by end users for entertainment purpose 4 1 AI-assisted Melody creation & Lyrics generation AI-assisted Orchestration, Instrumentation and structure 2 3 9 AI-assisted Enhancement/enrichment of existing/humancreated music (e.g., verses addition) 8 Gen AI economic impact - Use cases Music

53. 53 Study on the economic impact of Generative AI in the Music and Audiovisual industries Prioritisation of use cases based on expected impact on creators’ revenues and adoption probability – Matrix Analysis Prioritisation of Gen AI use cases in the Music field Source: PMP Strategy Analysis Automated complete musical output generation for commercial distribution (e.g. “mood playlists”) 5 Automated complete musical output generation for background content (in AV works, in-store sonorization, social media…) 7 AI-assisted Idea Generation: Complete/partial musical work creation (samples, moods, rhythms, etc.) 1 AI-assisted Enhancement/enrichment of existing/human-created music (e.g., verses addition) 8 AI-assisted Melody creation & Lyrics generation 2 AI-assisted Orchestration, Instrumentation and structure 3 AI-assisted Pitch correction, Editing, Mastering 9 Very Low Very High Adoption probability Very Low Impact on creators' revenues Very High Automated complete musical output generation for core content in audiovisual works (e.g., video games main theme) 6 Automated complete musical output generation for sole entertainment purpose (non-commercial use) 4 Main use cases High adoption Low impact Mid adoption Mid to low impact Mid Impact Low adoption Legend: use cases impact on creators’ revenues - Marginal - Moderate - High to very high Gen AI economic impact - Use cases 53 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music

54. 54 Study on the economic impact of Generative AI in the Music and Audiovisual industries Transformation of music streaming: end-users become music curators Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Technology & Output Quality: The tools are user-friendly and intuitive, but the quality of the outputs is still limited compared to traditional commercial music • Usage/Adoption: Usage remains primarily occasional and for entertainment purposes • Techno & quality of outputs: Improvement of the technology leading to increasingly higher-quality outputs • Usage/adoption: Widespread adoption and a shift from occasional, ad hoc use to regular use, similar to traditional streaming platforms With the perfecting and democratization of text-to-song and voice cloning tools, endconsumers are moving from simple users to music curators, thus questioning the very notion creator. Tech company Current application • Tools like Suno Audio, designed for the general public, allow users to generate and listen to music tracks created from a simple text prompt. • For now, these tools are mainly used on an ad hoc basis for entertainment purposes among friends, colleagues etc. 2028 potential application Two possible scenarios : 1. Tools like Suno and Udio evolve to become new players in the music streaming industry 2. Existing music streaming platforms integrate these new AI-powered content generation features themselves Example of service providers Example AI-powered platform allowing end-users to both produce and listen to AI-generated music (Suno app, see next page) 2028 est. level of adoption/maturity No direct impact but dilution of human-created tracks in the overall revenues of streaming 2 Revenue loss Direct revenues for Gen AI providers, either directly providing services, or internalized in DSPs Gen AI providers’ revenues 3 Market boost : monetization of these new features (for DSPs) or services (for AI tech companies) 1 Market size / Gen AI penetration Low Mid High Legend Automated complete musical output generation for sole entertainment purpose (non-commercial use) 4 2028 main economic impacts identified 2028 impact on creators’ revenues 54 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music Gen AI economic impact - Use cases

55. 55 Study on the economic impact of Generative AI in the Music and Audiovisual industries Suno launched its ‘Suno for mobile’ app in July 2024, offering enhanced functionalities Example - Suno’s latest app launch transforms passive music listening into an interactive experience Source: Suno.com website The Suno mobile app allows users to create and share music in new and innovative ways. The key functionalities of the app include: 1. Music Creation • Text-to-song : users can generate songs by inputting lyrics or descriptions • Audio recordings : The app allows the user to record an audio and use it for the song 2. Music Streaming • Music curation : The app provides tools to curate and collect music that the user enjoys from other creators ‘Suno for mobile’ promotion – Suno.com I suspect that the strategy of both Suno and Udio is to become new streaming platforms. […] where users can engage [with the content], create their own versions of it, republish it and become curators and creators themselves. – Music and AI expert Automated complete musical output generation for sole entertainment purpose (non-commercial use) 4 Gen AI economic impact - Use cases Music

56. 56 Study on the economic impact of Generative AI in the Music and Audiovisual industries 2028 est. level of adoption/maturity Penetration of AI-generated music on music streaming platforms Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Current tools can already produce good quality music for such purposes • Usage/adoption: Usage is still limited but the adoption remains difficult to quantify • Techno & quality of outputs: Improved technology with increasingly higher-quality outputs • Usage/adoption: high adoption potential for functional music and “passive” listening, for both individual and corporate customers Generative AI represents a potential opportunity for DSPs to generate royalty-free tracks and integrate them into their playlists. This approach could significantly boost their margins by drastically reducing copyright costs. Tech company Current application • Generative AI tools are already being used to create full music tracks for mainstream distribution on streaming platforms, often included in functional playlists. • AI-generated music could represent a significant portion of mainstream music, particularly in functional music and passive listening through suggested playlists (mood/contextual playlists) • DSPs might even use AI themselves to generate tracks, create and curate playlists based on user preferences and moods Automated complete musical output generation for commercial distribution (e.g. “mood playlists”) 5 Example AI-generated tracks produced to feed a DSP’s contextual “Morning Motivation” or “Casual Run” playlist Example of service providers 2028 potential application Low Mid High Legend 2028 Main economic impacts identified High potential cannibalisation of music creators’ streaming revenues 2 Revenue loss Revenues driven by subscription fees from prompters, or by the internalization into DSPs Gen AI providers’ revenues 3 Moderate penetration rate in volume and value on streaming platforms, particularly in mood playlists 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact - Use cases 56 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music

57. 57 Study on the economic impact of Generative AI in the Music and Audiovisual industries Spotify’s catalogue now includes AI-generated music created and uploaded by third parties Example – 100% AI-generated music is already streamed on DSPs Source: Spotify, Boomy, Expert interviews, Specialized press, Music Business Worldwide • AI-generated tracks are already circulating on streaming platforms, often featured in suggested playlists, with some generating substantial streams • Tools from third-party players such as Boomy facilitate the creation and upload of these Gen AI tracks on DSP platforms • The impact of this phenomenon, in terms of volume of tracks and streams, has yet to be quantified • This raises questions about how platforms should handle these tracks (whether they should be tagged for user identification and/or removed) • Managing this influx is challenging, as streaming services now receive about 1 million new songs each week Example of AI-generated songs on Spotify Boomy is a platform allowing the creation of AI-generated music to be uploaded on DSPs Boomy allows users to : 1. Create and edit songs 2. Release created music on DSPs 3. Use the musical work for : • Non-commercial purposes in video, livestreaming, and other songs • Commercial purposes in podcasts and social media and social media advertising 1 2 3 • In 2023, Boomy had created 14.4 million songs. • The platform retains the copyright for all songs created, while users receive an 80% share of the royalty distribution fees Gen AI economic impact - Use cases Automated complete musical output generation for commercial distribution (e.g. “mood playlists”) 5 Music

58. 58 Study on the economic impact of Generative AI in the Music and Audiovisual industries Study on the economic impact of Generative AI in the Music and Audiovisual industries 2028 est. level of adoption/maturity Rise of tailored AI-generated music for social media content Current level of adoption/maturity • Techno & quality of outputs: Current services can produce quality musical content for such purposes • Usage/adoption: Mass use remains limited as these tools are not fully integrated within major social networks functionalities Thanks to Generative AI, content creators can quickly create royalty-free music perfectly suited to their YouTube videos for instance CMO Current application • Music generation for social media content using Gen AI-powered tools is already underway • Platforms are already investing in this technology (e.g., TikTok, with the acquisition of Jukedeck in 2019) or developing their own tools (e.g., Meta with AudioCraft) • In addition to the current prompt-to-music system, Gen AI will allow to provide instant, context-aware music for user-generated content across all social media platforms • Platforms will continue to invest in and promote Gen AI (copyright-free) music in their music libraries for content creators Example A TikTok video with AI-generated music specifically tailored to the video’s content 2028 potential application Source: Experts’ interviews, Specialized press, PMPS Scenario Planning • Techno & quality of outputs: Improved technology with increasingly higher-quality outputs, with tools that are already user-friendly, easy to use, and feature advanced UX • Usage/adoption: Widespread use can be expected with the integration of AI music generation tools into major social networks Low Mid High Legend (3) Generalist players Specialized players Internal tools Example of service providers AudioCraft by Automated complete musical output generation for background content (AV works, in-store, social media…) 7 Main economic impacts identified High potential cannibalisation of a portion of music creators’ social media revenues 2 Revenue loss Revenues driven by B2B subscription fees and direct orders (brands, content creators…) Gen AI providers’ revenues 3 Very high penetration rate of Gen AI outputs in user-generated content on social media 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact - Use cases Music

59. 59 Study on the economic impact of Generative AI in the Music and Audiovisual industries Study on the economic impact of Generative AI in the Music and Audiovisual industries Widespread adoption of Gen AI music for background content in audiovisual works or public places Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Current tools can already produce good quality music for such purposes • Usage/adoption: Penetration rates remain limited so far Generative AI has the potential to significantly impact background music, particularly in tasks where high volumes and quick production times are key, much like traditional music libraries. CMO Current application • Gen AI tools are already used to generate background scores for various projects and applications (e.g., advertising, sound systems in public places etc.) but remain limited so far • In addition to the use for background scores, Gen AI could provide customizable and context-sensitive background music services for a wide range of multimedia Automated complete musical output generation for background content (AV works, in-store, social media…) 7 Example of service providers Example AI-generated jingle for a TV show 2028 potential application 2028 est. level of adoption/maturity • Techno & quality of outputs: Improved technology with increasingly higher-quality outputs • Usage/adoption: High potential for adoption by B2B clients to reduce costs Low Mid High Legend Main economic impacts identified High cannibalisation rates: replacement of human produced “production music” for B2B use 2 Revenue loss Gen AI tools’ revenues driven by B2B subscription fees and direct orders Gen AI providers’ revenues 3 Very high penetration rate of Gen AI outputs in the music library market segment 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact - Use cases Music

60. 60 Study on the economic impact of Generative AI in the Music and Audiovisual industries Example - Meta invests in AI to improve its music library and enhance content creators experience Source: Photo credit : Social Media Examiner Meta’s free access music library for social media posts Meta has been active in the library music market since 2017: • The Meta Sound Collection offers a library of Meta-owned audio clips available for free In parallel, Meta invests in AI to expand its collection of copyright-free music content : • In 2023, Meta launched AudioCraft, an open-source AI model allowing the generation of high-quality, realistic audio and music from text-based user inputs Automated complete musical output generation for background content (AV works, in-store, social media…) 7 Gen AI economic impact - Use cases Music

61. 61 Study on the economic impact of Generative AI in the Music and Audiovisual industries Study on the economic impact of Generative AI in the Music and Audiovisual industries Example of service providers Moderate penetration of Gen AI music for core content in audiovisual works Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Current services can produce quality musical content, but they are not always considered sufficient to fully replace high-impact and high-budget commissioned creations • Usage/adoption: Use remains limited • Techno & quality of outputs: Gen AI services will rapidly be able to offer highly qualitative content • Usage/adoption: Expected to become a more widely used tool in audiovisual production for all application types such as series, movies, video games etc., except for high-budget projects requiring the support of famous industry names Generative AI enables advanced personalisation and real-time music creation. In video games for instance, it can generate continuous music streams in a specific style, adapting the sound to match in-game events. Tech Company Current application • Very limited applications today, with substantial musical content for audiovisual works remaining mostly commissioned compositions • Gen AI is used to generate outputs replacing commissioned works for core content in certain audiovisual works (lower production budget) • In addition, Gen AI could be used for more advanced personalization and real-time music creation (e.g., video games ) Automated complete musical output generation for core content in audiovisual works (e.g., video games main theme) 6 Example AI-generated substantial musical soundtrack for a video game production 2028 potential application 2028 est. level of adoption/maturity Low Mid High Legend Main economic impacts identified High potential cannibalisation of revenues for audiovisual music composers (less orders / commissioned works) 2 Revenue loss Revenues driven by B2B subscription fees and direct orders mainly from B2B producers Gen AI providers’ revenues 3 High penetration rate of Gen AI outputs for background music in AV works, for cost reduction 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact - Use cases Music

62. 62 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approach and Methodology What will be the economic impact of Generative AI in the Music field by 2028? Main Applications What will be the economic impact of Generative AI in Audiovisual by 2028? 2028 forecast Economic impact in Music creation

63. 63 Study on the economic impact of Generative AI in the Music and Audiovisual industries Projected evolution of Gen AI music outputs market size | €Bn, 2023 - 2028 Fully Gen AI outputs in Music are expected to be worth c.€16Bn in 2028, doubling on average each year Source: PMP Strategy analysis 2023 2024 2025 2026 2027 2028 €1Bn €7Bn €11Bn Note: In this market size calculation, no distinction is made whether the music outputs are copyrightable or not | (1) 2023-2028 CAGR +100% avg. per year(1) Gen AI based music streaming Gen AI music in traditional streaming Gen AI music for B2B purposes 0.5 €Bn 4 €Bn 16 €Bn 1 Market size Gen AI Music economic impact - Forecast

64. 64 Study on the economic impact of Generative AI in the Music and Audiovisual industries Fully Gen AI music outputs market size | €Bn, 2028 This market will be mostly driven by Gen AI music on streaming platforms and Gen AI music for B2B purposes Source: PMP Strategy analysis Note: In this market size calculation, no distinction is made whether the music outputs are copyrightable or not €1Bn €10Bn €5Bn Gen AI based music streaming Gen AI music in traditional streaming Gen AI music for B2B purposes €16Bn With the perfecting and democratization of prompt-to-”songs” (outputs), endconsumers are moving from simple users to music curators, thus questioning the very notion of the creator. New “streaming” possibilities will likely emerge from this. Tech company Generative AI represents a potential opportunity for DSPs to generate royalty-free tracks and integrate them into their playlists. This approach could significantly boost their margins by drastically reducing copyright costs. Tech company Gen AI platforms like Boomy are disrupting the distribution system. They offer prompter users the chance to directly distribute their creations (outputs) directly on Spotify. CMO Generative AI has the potential to significantly impact background music, particularly in tasks where high volumes and quick production times are key, much like traditional music libraries. Tech company 1 Market size Gen AI Music economic impact - Forecast

65. 65 Study on the economic impact of Generative AI in the Music and Audiovisual industries Market saturation will be significant on the music library segment (c.60%), with B2B clients looking to reduce costs Source: PMP Strategy analysis Gen AI based music curation and streaming 100% Gen AI music in traditional streaming platforms 20% Gen AI music for B2B purposes (audiovisual, social media, public places sound system, brands) 57% • New services allowing both music listening and curation/creation (see associated use case) • 2 potential scenarios: (i) Tools like Suno and Udio become new players in the music streaming industry, (ii) Existing music streaming platforms integrate and monetize these new Gen AI features • Penetration of Gen AI outputs in DSPs catalogues, generated by third parties or by the streaming platforms themselves (see associated use case) • High potential in mood music playlists (e.g., “Morning Coffee”, “Beast Mode” on Spotify) where end-users adopt a more passive listening • Use of fully Gen AI music for background music in audiovisual works, advertising, social media, in-store sonorization (see associated use cases) • High adoption rates fostered by B2B clients looking for costs reductions and the unlimited potential of these Gen AI outputs for B2B consumers (brands, audiovisual professionals, content creators, etc…) Market segments Fully Gen AI penetration rate in 2028(1) Use cases supporting the penetration of Gen AI outputs Note: (1) Weighed penetration rates of subcategories analysed 1 Market size Gen AI Music economic impact - Forecast

66. 66 Study on the economic impact of Generative AI in the Music and Audiovisual industries 99% 1% 5% 94% <1% 18% 80% 3% €36 Bn €43 Bn Music streaming B2C(1) revenues, generated by human created works vs. Gen AI outputs | €Bn, 2023 - 2028 A Gen AI boost is expected on the music streaming segment, due to new usage and functionalities which will be monetized by traditional or new players Source: PMP Strategy analysis €57 Bn Gen AI outputs penetrate the market in the traditional music streaming platforms and in new Gen AI based streaming services In 2028, Gen AI outputs could create an estimated additional €1.4Bn in the market, but the major part (c.€10Bn) of the revenue generated will be a substitution of humancreated works’ revenues In the current market, 99% of the € 36Bn music streaming market is generated by human creations Gen AI boost – additional value (2) Gen AI music in streaming – replacing human music Human-created works in streaming Human-created works in streaming Gen AI music in streaming 2023 2025 2028 Note: (1) Music streaming platforms B2B revenues (brands, in-store sonorization…) have been excluded of this analysis (included in the B2B segment of the market size calculation) | (2) Gen AI based music platforms (new Gen AI platforms or new offers of traditional DSPs) 1 Market size Gen AI Music economic impact - Forecast

67. 67 Study on the economic impact of Generative AI in the Music and Audiovisual industries Revenues for Music creators with and without the impact of Generative AI | €Bn and %, 2023 - 2028 Under current conditions, this market penetration by Gen AI outputs could put 24% of Music creators’ revenues at risk by 2028 Source: PMP Strategy analysis, CISAC global collections report Note: In this analysis, creators' revenues are represented by CMOs collections Share of creators’ revenue at risk due to Gen AI substitution Revenues for creators with the impact of Gen AI 2028 total cannibalisation % on total revenues €4Bn 24% ∑= €10 Bn 10 16 2023 2024 2025 2026 2027 2028 ’23-’28 cumulated loss Revenue 2 loss Gen AI Music economic impact - Forecast

68. 68 Study on the economic impact of Generative AI in the Music and Audiovisual industries The potential impact will be strong on Digital collections (up to 30% cannibalisation), TV & Radio and Background (c. 22% of collections) Source: PMP Strategy analysis Note: (1) Cannibalisation rates estimated based on interviews and workshops with CMOs and industry experts Gen AI cannibalisation rate in 2028(1) Use cases explaining cannibalisation levels Music creators’ revenue streams Digital TV & Radio(2) Live & Background CD & Video Other streams 30% 22% 22% 21% • Gen AI music in social media content, audiovisual streaming (AV works’ background music), music streaming (mainly in mood music) and video games • Gen AI music for background content in audiovisual works (mainly in lower production budget works) • Gen AI music for the sonorisation of public places (stores, malls, restaurants; using personalised and unlimited Gen AI outputs to reduce costs) • Gen AI music in video games (use case with high level of adoption, due to new unlimited tracks for games – by moods, styles – enabled by Gen AI) Marginal to no economic impact Revenue 2 loss Gen AI Music economic impact - Forecast

69. 69 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music creators’ revenue share in the music streaming market | €Bn, 2023 - 2028 In a growing music streaming market, creators’ share will thus decrease further due to Gen AI (-1.5pts) Source: PMP Strategy analysis, CISAC global collections report, • In 2023, the share of creators' revenue in the streaming market amounts to approximately 8% • In 2028, this share could decrease to c.6%, on a significantly higher market • This dilution could represent a loss of c.€0.9Bn for creators in 2028 and a cumulated loss of c.€2.3Bn in the five coming years Note: (1) Music streaming platforms B2B revenues (brands, in-store sonorization…) have been excluded of this analysis (included in the B2B segment in the market size calculation Revenue for creators without Gen AI Revenue for creators with Gen AI 2023 2024 2025 2026 2027 2028 5% 10% 8.2% 8.2% 8.0% 8.2% 7.5% 8.2% 7.2% 8.2% 6.8% 8.2% 6.4% c.0.9Bn in revenue loss for music creators in streaming only Revenue 2 loss Gen AI economic impact - Forecast Music

70. 70 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music Gen AI providers’ revenues | €Bn, 2023 - 2028 Gen AI providers’ revenues in Music could reach c.€4Bn in 2028, doubling on average each year from 2023 to 2028 Source: PMP Strategy analysis 2023 2024 2025 2026 2027 €1Bn €2Bn 2028 €0.3Bn €1Bn €2Bn Mass public tools for complete outputs’ generation Gen AI softwares for assistance in the creative process c.0.1 €Bn c4 €Bn c.0.6 €Bn Note: (1) 2023-2028 CAGR Gen AI services 3 revenues Gen AI Music economic impact - Forecast +100% avg. per year(1)

71. 71 Study on the economic impact of Generative AI in the Music and Audiovisual industries Source: PMP Strategy Gen AI outputs in Music will be worth a cumulative €40Bn over the next five years, rising to an annual value of €16Bn in 2028 By 2028, Gen AI music will account for around 20% of traditional music streaming platforms’ revenues and around 60% of music libraries revenues Market size 1 €16Bn Estimated market value of Gen AI outputs in Music in 2028 Under current conditions, this market penetration by Gen AI outputs could put 24% of Music creators’ revenues at risk in 2028 This represents a cumulative loss of €10Bn over the next 5 years, and an annual loss of €4Bn in 2028 Revenue loss 2 €4Bn|24% Creators' revenues at risk in 2028 compared to a no Gen AI situation Gen AI services’ revenues 2 €4Bn Estimated revenues of Gen AI Music services in 2028 Gen AI services are projected to generate exponential revenue growth, reaching an estimated €4Bn in 2028, with a cumulative total of €8Bn over 5 years Study key takeaways – Music Gen AI economic impact - Forecast Music

72. 72 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approach and Methodology What will be the economic impact of Generative AI in the Music field by 2028? What will be the economic impact of Generative AI in Audiovisual by 2028? Main Applications 2028 forecast Economic impact in Audiovisual creation

73. 73 Study on the economic impact of Generative AI in the Music and Audiovisual industries Identification of Gen AI main applications in the audiovisual creation process Source: Experts’ interviews, specialized press Ai-assisted video editing, mastering, voice removing, visual effects… Dubbing-subtitling Automated translation and adaptation AI-assisted storyboard generation, scene visualization Creative inspiration / Research and planning Creation / Execution Post-production Main use case AI-assisted or automated complete audiovisual outputs creation/composition AI-assisted post-production AI-assisted video & audiovisual workconceptualization/idea generation 7 Automated complete audiovisual output generation for higher production budget works (series, films…) Automated complete audiovisual output generation for entertainment purpose (mainly short form) 3 1 8 Screenwriting Automated screenwriting 2 Automated complete audiovisual output generation for lower production budget works (ads, children’s cartoons, ..) 4 AI-assisted or automated complete audiovisual works generation for social media content (short form, mainly user generated content) 5 6 Gen AI economic impact – Use cases Audiovisual

74. 74 Study on the economic impact of Generative AI in the Music and Audiovisual industries Prioisation of use cases based on expected impact on creators’ revenues and adoption probability – Matrix Analysis Prioisation of Gen AI use cases in the Audiovisual field Source: PMP Strategy Analysis Very Low Very High Adoption probability Very Low AI-assisted Video editing, Mastering, Voice removing, Visual Effects… 8 AI-assisted Storyboard generation, Scene visualization 1 Automated complete audiovisual output generation for higher production budget works (series, films…) 5 Automated complete audiovisual output generation for entertainment purpose (mainly short form) 3 Automated translation and adaptation 7 AI-assisted or automated complete audiovisual works generation for social media content (short form, mainly user generated content) 6 Main use cases Mid to High adoption Low impact High impact Low adoption Automated screenwriting 2 Automated complete audiovisual output generation for lower production budget works (ads, children’s cartoons, ..) 4 Impact on creators’ revenues Very High Gen AI economic impact – Use cases Legend: use cases impact on creators’ revenues - Marginal - Moderate - High to very high Audiovisual

75. 75 Study on the economic impact of Generative AI in the Music and Audiovisual industries Example of service providers Widespread adoption of Gen AI tools for video content generation on social media Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Currently, AI services cannot produce complex and highly customized visuals using the prompt-to-video method • Currently, adoption is low, as the technology is still developing • With expected strong improvement in technology capacities, by 2028, the adoption is expected to be very high, with AI-assisted tools becoming standard in the content creation process, particularly for creators who need to produce high-quality, visually rich content on tight deadlines Current application • Generation of illustrative, moderate quality videos that support and enhance various types of content • Automating tasks like adding relevant visuals, animations, and effects, allowing them to produce engaging content more efficiently • Generation of high quality, longer video sequences with minimal input Example AI-assisted or automated complete audiovisual works generation for social media content (short form, mainly user generated content) 6 Illustrative videos supporting a history-themed video on YouTube Make-a-Video by 2028 potential application 2028 est. level of adoption/maturity Low Mid High Legend Gen AI will democratize creation and increase user-generated audiovisual content. If technology allows it, we'll have tools enabling to create affordable, high-volume short videos for social networks which will flood the market. Audiovisual institution Main economic impacts identified High cannibalisation of creators’ revenues from traditional video production services 2 Revenue loss Revenues driven by subscription fees from content creators and other social media users Gen AI providers’ revenues 3 High penetration rate and market growth with the use of Gen AI video mainly for content creators 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact – Use cases Audiovisual

76. 76 Study on the economic impact of Generative AI in the Music and Audiovisual industries Gen AI automated subtitling for a documentary production Improving technology for the automation of translations and adaptations Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Currently, AI services can produce complex and highly customized visuals using the prompt-to-video method • Currently, adoption is low for complex and high-quality audiovisual works but is high for lower value content • By 2028, adoption is expected to be very high, with automated dubbing and subtitling becoming more qualitative and therefore a standard practices in the industry, particularly for streaming platforms and global content distributors. If the technology improves, Gen AI could be used to produce high-quality content. However, as of now, automated subtitling and dubbing are still limited to low added-value applications. Dubbing and Subtitling Agency Current application • Automating speech translation, subtitles synchronisation and voice dubbing generation • Application is currently focused on non-substantial video due to technology capacity • High-quality real-time translations and lip-syncing that are indistinguishable from human performance Example 7 Automated translating/adapting Example of service providers 2028 potential application 2028 est. level of adoption/maturity Low Mid High Legend Main economic impacts identified High cannibalisation of revenues for DB/ST authors and less orders for DB/ST agencies 2 Revenue loss Revenues driven by subscription fees from audiovisual producers for DB/ST works Gen AI providers’ revenues 3 High penetration rate of Gen AI DB/ST; potential shrinking of the overall market (cost reduction) 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact – Use cases Audiovisual

77. 77 Study on the economic impact of Generative AI in the Music and Audiovisual industries Fully Gen AI automated screenwriting for a TV soap opera Generative AI as an assistant and/or a substitution for screenwriting in audiovisual works Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: easy to use tools • Currently, adoption is still pretty low as tools are mostly used for assistance / idea generation or rewriting • By 2028, adoption is expected to be very high especially on lower production value segments, with AI tools allowing to create ever more complex stories based on specified criteria The main concerns about the impact of Gen AI often arise from screenwriters among our members: the profession is likely to be heavily impacted in the coming years. Audiovisual CMO Current application • Current tools allow to automate / facilitate a number of tasks related to scriptwriting: scenario analysis, research, rewriting, … • The progress of tools will allow to generate more quality scripts and automate the full generation of scenarios for certain contents Example Example of service providers 2028 potential application 2028 est. level of adoption/maturity Low Mid High Legend Main economic impacts identified High cannibalisation of revenues for screenwriters as Gen AI scripts become more cost-effective 2 Revenue loss Revenues driven by subscription fees from audiovisual producers for screenwriting works Gen AI providers’ revenues 3 High penetration rate of Gen AI scripts; potential shrinking of the overall market (cost reduction) 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact – Use cases 2 Automated screenwriting Audiovisual

78. 78 Study on the economic impact of Generative AI in the Music and Audiovisual industries Example of service providers Rise of Gen AI content for lower budget audiovisual productions, fostered by producers’ willingness to gain efficiency Source: Experts’ interviews, Specialized press, PMPS Scenario Planning Current level of adoption/maturity • Techno & quality of outputs: Currently, AI services cannot produce quality audiovisual content • Currently, adoption is low, due to the poor quality of AI-automated audiovisual outputs generation • Mid- to high adoption, depending on the evolution of the quality of the audiovisual content/output Current application • First creations of entire audiovisual outputs in animated fiction films & series • Ads or Music clips generation, with enhanced possibilities but mid production quality • High-quality audiovisual content that is indistinguishable from human-made productions • Potential seamless integration into some audiovisual production sectors (soap operas, advertising, music clips, animated works…) Automated complete audiovisual output generation for lower production budget works (ads, soap opera...) 4 Example Extract of Qianqiu Shining, China Media Group AI-generated animated series in 2022 Adoption depends on the technology's capabilities. Currently, producing a high-quality audiovisual work from start to finish, especially a two-hour film, is beyond reach, while it is already feasible for a short-animated film. Audiovisual institution 2028 potential application 2028 est. level of adoption/maturity Low Mid High Legend Main economic impacts identified Very high impact on audiovisual creators/authors due to productions’ budget decrease 2 Revenue loss Revenues driven by B2B subscription fees from audiovisual production companies Gen AI providers’ revenues 3 High penetration rate of Gen AI; shrinking of the traditional market (cost reduction) 1 Market size / Gen AI penetration Expected impact on creators’ revenues Gen AI economic impact – Use cases Audiovisual

79. 79 Study on the economic impact of Generative AI in the Music and Audiovisual industries Approach and Methodology What will be the economic impact of Generative AI in the Music field by 2028? What will be the economic impact of Generative AI in Audiovisual by 2028? Main Applications 2028 forecast Economic impact in Audiovisual creation

80. 80 Study on the economic impact of Generative AI in the Music and Audiovisual industries Projected evolution of Gen AI audiovisual outputs market size | €Bn, 2023 - 2028 Fully Gen AI audiovisual outputs are expected to be worth c.€48Bn in 2028, with an average growth of c.85% each year Source: PMP Strategy analysis 2023 2024 2025 2026 2027 2028 €6Bn €22Bn €34Bn Note: In this market size calculation, no distinction is made whether the Gen AI outputs are copyrightable or not, and only Fully Gen AI audiovisual/video outputs are considered | (1) 2023-2028 CAGR AVOD (incl. Youtube) Video streaming (SVOD & downloads) TV broadcasters Social media 13 €Bn 48 €Bn 2 €Bn +85% avg. per year(1) Gen AI economic impact - Forecast 1 Market size Audiovisual

81. 81 Study on the economic impact of Generative AI in the Music and Audiovisual industries Fully Gen AI audiovisual outputs’ market size | €Bn, 2028 Source: PMP Strategy analysis Note: In this market size calculation, no distinction is made whether the Gen AI outputs are copyrightable or not | (1) includes SVOD, FAST, Pay-per-view, EST (downloads) €11Bn €3Bn €5Bn €30Bn TV broadcasters SVOD AVOD (incl. Youtube) Social media €48Bn Some specific YouTube creators now have the potential to produce fully Gen AI videos, enabling them to deliver high-quality, engaging content on complex topics with minimal effort. For instance, a channel dedicated to science education can leverage Gen AI to generate fully automated content. Audiovisual institution There is a risk that some audiovisual production with lower budget might be replaced in the future by Gen AI outputs, for example some soap operas, low-budget advertising & clips, or animated content for kids… Producers of these content will be looking to reduce costs by leveraging Gen AI tools. CMO Gen AI will democratize creation and increase user-generated audiovisual content. If technology allows it, we'll have tools enabling to creating affordable, high-volume short videos for social networks which will flood the market. Audiovisual institution (1) The market will be mostly driven by the penetration of Gen AI outputs on social media and in lower production value TV programmes Gen AI economic impact - Forecast 1 Market size Audiovisual

82. 82 Study on the economic impact of Generative AI in the Music and Audiovisual industries Market saturation by Gen AI complete audiovisual outputs will remain more limited than for Music Source: PMP Strategy analysis 4% 2% TV Broadcasters SVOD AVOD 8% • Leveraging of Gen AI content to produce lower budget and more personalized programmes (lower budget TV shows, advertisings, kids animated, …) allowing to keep up with audience demand at reduced costs • Penetration of Gen AI audiovisual outputs in SVOD platforms catalogues • Use of Gen AI tools to produce audiovisual outputs for short and viral content on AVOD platforms (mainly YouTube) Market segments Fully Gen AI penetration rate in 2028(1) Use cases supporting the penetration of Gen AI outputs Note: (1) Weighed penetration rates of subcategories analysed Social media 13% • Use of Gen AI tools to automate the generation of short form videos on social media, and produce engaging and personalized content more efficiently by adding relevant visuals, animations, and effects 1 Market size Gen AI economic impact - Forecast Audiovisual

83. 83 Study on the economic impact of Generative AI in the Music and Audiovisual industries Revenues for Audiovisual creators with and without the impact of Gen AI outputs | €Bn and %, 2023 - 2028 The use of Gen AI tools to automate tasks in the production process could put 21% of audiovisual creators’ revenue at risk by 2028 Source: PMP Strategy analysis, CISAC global collections report 2028 total cannibalisation % on total revenues €4.5Bn 21% ∑= €12 Bn Note: In this analysis, revenues include both CMOs collections and other revenue streams (upfront payments) 10 22 2023 2024 2025 2026 2027 2028 Share of creators’ revenue at risk (due to substitution of complete Gen AI outputs and use of Gen AI tools in the production process) Revenues for creators with the impact of Gen AI ’23-’28 cumulated loss Gen AI economic impact - Forecast Revenue 2 loss Audiovisual

84. 84 Study on the economic impact of Generative AI in the Music and Audiovisual industries The potential impact will be particularly strong for Translators and Adapters (c.56% of cannibalisation rate) Source: PMP Strategy analysis Note: (1) Cannibalisation rates estimated based on interviews and workshops with CMOs and industry experts Directors & other co-authors Translators / Adapters 15% 20% 56% • Widespread use of Gen AI tools to automate directors and other co-authors’ tasks, fostered by producers’ willingness to gain efficiency and reduce costs • Complete audiovisual outputs replacing human-created works on certain categories of audiovisual content • Use of Gen AI tools for automatic translation and adaptation, with outputs increasingly closer to human work at a decreasing cost Gen AI cannibalisation rate in 2028(1) Use cases explaining cannibalisation levels Audiovisual creators / authors’ categories Screenwriters • Use of Gen AI screenwriting assistance tools, supporting authors in their work but also pushing producers to reduce the budget spend for screenwriting Revenue 2 loss Gen AI economic impact - Forecast 84 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual

85. 85 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual Gen AI providers’ revenues | €Bn, 2023 - 2028 In Audiovisual, Generative AI providers’ revenues could reach €5Bn in 2028, driven by Gen AI prompt-to-outputs tools Source: PMP Strategy market model, Experts’ interviews 2023 2024 2025 2026 2027 67% 33% 2028 €0.5Bn €2Bn €4Bn Audiovisual prompt to outputs services market size Gen AI-assisted video/audiovisual creation softwares c.0.2 €Bn c.5 €Bn c.1 €Bn Gen AI services 3 revenues +85% avg. per year(1) Note: (1) 2023-2028 CAGR Gen AI economic impact - Forecast Audiovisual

86. 86 Study on the economic impact of Generative AI in the Music and Audiovisual industries Source: PMP Strategy AI-generated complete Audiovisual outputs are expected to be worth c. €48Bn in 2028. Audiovisual outputs generation for social media and TV will account for the lion’s share of the market. Market size 1 €48Bn Estimated market value of Gen AI outputs in Audiovisual in 2028 The widespread use of Gen AI tools throughout the production process of audiovisual works could put 21% of creators' revenue at risk by 2028. This represents a cumulative loss of €12Bn over the next 5 years, and an annual loss of €4.5Bn in 2028. Revenue loss 2 Gen AI services’ revenues 2 €5Bn Estimated revenues of Gen AI Audiovisual services in 2028 Gen AI services in Audiovisual (both mass public and professional tools/softwares) are projected to generate exponential revenue growth, reaching an estimated €5Bn in 2028, with a cumulative total of €13Bn over 5 years. Study key takeaways – Audiovisual €4.5Bn|21% Audiovisual creators' revenues at risk in 2028 (compared to a no Gen AI situation) Gen AI Audiovisual economic impact - Forecast

87. 87 Study on the economic impact of Generative AI in the Music and Audiovisual industries Detailed methodology and assumptions – Music Glossary Detailed list of interviews conducted PMP Strategy presentation Appendix

88. 88 Study on the economic impact of Generative AI in the Music and Audiovisual industries Market segmentation and impact of Gen AI use cases done by crossing 3 dimensions Use cases analysis has allowed to identify the Music market segments most likely to be impacted by Gen AI in the next 5 years Source: PMP Strategy methodology Legend: expected impact of Gen AI outputs on the market segment (market penetration of Gen AI outputs) - Marginal to low - Moderate - High to very high Production budget of the works Higher budget works Lower budget works Type of musical artwork Commercial music Mood music Background music/sound effects (for social media, audiovisual, public places) Core music in audiovisual content Music works distribution channel Digital – current streaming platforms Digital – New AI-based streaming platforms Digital – other downloads Physical sales (CDs, vinyls...) Music libraries Direct commissioned works B2C / B2B2C: B2B – Music for audiovisual, social media, in-store sonorization, brands…: Concerts/Festivals/Music shows (Live audience) 1 Market size Appendix – Detailed methodology and assumptions Music

89. 89 Study on the economic impact of Generative AI in the Music and Audiovisual industries For each of these segments, a Gen AI penetration rate has been estimated based on use cases expected adoption and impact (from low to very high) Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews 2023 market size and 2028 forecasts on segments impacted by Gen AI Fully Gen AI outputs Penetration rate in 2028 2028 Fully Gen AI music outputs’ market size 1 Market size x = ‘23 market size ’28 forecasts c. €16Bn New Gen AI music streaming platforms(1) Music streaming platforms <0.1Bn €36Bn €5Bn 1.4Bn €55Bn €7.8Bn • New Gen AI based streaming platforms allowing listeners/users to curate and listen to Gen AI music and/or new offers in current streaming platforms • Commercial music • Mood music • Licensed music – commercial/preexisting tracks • Licensed music – background/sound effects • “Royalty free” music (buy out) 100% 20% 57% €1.4Bn €10Bn €4.5Bn Total Gen AI market size for musical outputs • We consider here the “public price” of music for B2B clients buying / commissioning musical works for audiovisual content, public places sonorization, etc. • The impact on the value driven by the diffusion of this music in audiovisual works (by SVOD platforms, TV/Radio broadcasters) is considered as part of question 2/ • The market size calculation includes royalty-free music content Note: (1) Does not include all AI-based music services, but only new platforms on which end-users/subscribers are both music listeners and creators (Udio, Suno) | (2) Also includes music libraries from majors (e.g. Universal Music Production), and from DSPs B2C/B2BC2 B2B Music libraries(1) - for audiovisual, social media, public places sonorization, brands Weighed penetration rates considering : Appendix – Detailed methodology and assumptions 89 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music

90. 90 Study on the economic impact of Generative AI in the Music and Audiovisual industries Note: (1) Segmentation of this category crossing the type of audiovisual content on which it is played and the importance of the music in the AV work Music market segmentation on segments impacted by AI generated outputs Revenue cannibalisation for Music creators has been calculated based on a segmentation of the global CMOs collections Source: PMP Strategy market model, Experts’ interviews Revenue loss 2 • Segmentation approach to measure AI outputs’ cannibalisation on creators’ revenues: – Breakdown of CISAC 2023 global Music collections by categories (e.g Digital) and subsegments (e.g Music streaming platforms, SVOD platforms, Social networks, etc.) – 2028 collections forecast for each subsegment based on historical growth rates and future market trends – For each subsegment: estimated cannibalisation rate in 2028 based on use cases and market estimates conducted as part of question 1/ Key methodology insights Digital CISAC Global collections Collections segmentation by use Rationale • Exclusion from the calculation (blurred boxes) of the collections’ subsegments where the human created works are unlikely to be replaced by AI: • E.g. : it is estimated that the music used for live shows will remain predominantly human-created, as the audience always associates itself with a musician when attending a live event TV & Radio(1) CD & Video synchronisation Others Live & Background Music streaming AV streaming Social networks Video Games, Download & other digital Original music of high production value AV work Background music for high production value AV works Original music of low production value AV work Background music of low production value AV wors Live audience performances Background music Music as main focus in public places Other sources CD, Vinyls DVD, Blue rays Video games Appendix – Detailed methodology and assumptions Music

91. 91 Study on the economic impact of Generative AI in the Music and Audiovisual industries CISAC Music collections evolution by music usage in current market evolutions | €Bn, 2018-2028 Based on historical growth and future market trends, CISAC collections for Music are expected to reach c. €15.8Bn by 2028 Source: PMP Strategy market model, Experts’ interviews Forecasts 0% 19% 4% 2018 2019 2020 2021 2022 39% 29% 26% 3% 0% 3% 2023 2024 2025 2026 2027 44% 22% 27% 3% 39% 1% 30% 3% 2028 8.5 8% 8.2 8.5 10.8 11.8 8.4 13.2 14.0 14.9 15.8 12.4 Digital TV & Radio Live & Background CD & Video Synchronisation Others YoY growth (%) -1% -2% +4% +28% +9% +6% +6% +6% +6% +6% CAGR 18-22 22-24 +6% +7% +26% +10% +2% -2% 24-28 +6% +9% +1% -1% +14% +7% -15% +6% +2% +13% +14% +10% -2% +13% +6% Post-Covid recovery 15.8 €Bn Note: Forecasts CAGR based on historical collections and upcoming trends Revenue loss 2 Appendix – Detailed methodology and assumptions Music

92. 92 Study on the economic impact of Generative AI in the Music and Audiovisual industries To estimate the revenue loss for Music creators, key assumptions have been made on Gen AI cannibalisation rates by collections subsegments Source: PMP Strategy market model, CISAC collections, Specialized press, Experts’ interviews Revenue loss 2 ‘23 collections ’28 forecasts c. €3.8Bn (24%) Total revenue loss for music artists in 2028 x = Revenue sub-streams (1), 2023 collections and 2028 forecasts 2028 revenue loss for creators Digital €4.5Bn • Music streaming • AV streaming • Social media • Video Games €6.9Bn 30% €2Bn TV & Radio(1) €3.4Bn • Original music for higher budget AV works/ music tracks on radio • Background music for higher budget AV works • Original music for lower budget AV works • Background music for lower budget AV works €3.5Bn 22% €0.7Bn Live & Background €3.1Bn • Live audience performances • Music as main focus in a live audience (clubs…) • Background music (sonorization of public places) €4.2Bn 21% €0.8Bn CD & Video €0.4Bn • CD, Vinyls • DVD, Blue-rays • Video games €0.4Bn 21% €0.1Bn Weighed cannibalisation rates considering : Gen AI outputs 2028 cannibalisation rates Note: (1) AV works in the cannibalisation rates consider both TV & radio works Other substreams €0.3Bn €0.7Bn Total collections €11.8Bn €15.8Bn Marginal to no impact Appendix – Detailed methodology and assumptions Music

93. 93 Study on the economic impact of Generative AI in the Music and Audiovisual industries Music Gen AI tools providers’ revenues – 2023-2028, €Bn Gen AI tools providers have been segmented in 2 categories to assess their revenues Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights 1. Mass public tools 2. Gen AI tools/software as an assistant Number of Gen AI based services in 2028 Average Number of users in 2028 and % of paying subscriptions Average annual price of the subscription €1.4Bn(1) 2028 associated players total revenues 4 60m (10%) €60 x x = Audio/Music software market size 2023 2028 €2.3Bn €4.2Bn(2) 2028 Generative AI tools penetration rate €2.3Bn 2028 associated players = total revenues 60% / High x Note: (1) Calculated in question 1, representing new AI based streaming platforms (and described as the use case: end-user as curators) | (2) High growth rate as new services (providing both prompt-to-outputs and assistance in the creation – such as Boomy and Beatoven) will lead to revenue growth on the market Gen AI services 3 revenues Appendix – Detailed methodology and assumptions Music

94. 94 Study on the economic impact of Generative AI in the Music and Audiovisual industries Detailed methodology and assumptions – Audiovisual Glossary Detailed list of interviews conducted PMP Strategy presentation Appendix

95. 95 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual market segmentation(1) and impact of Gen AI use cases done by crossing 3 dimensions Use cases analysis has allowed to identify the Audiovisual market segments most likely to be impacted by Gen AI in the next 5 years Source: PMP Strategy methodology - Marginal - Moderate - High to very high Artistic value of the works / Nature of the programmes Type of audiovisual content Audiovisual works distribution channels/media Cinema TV broadcasters Radio broadcasters Digital - AVOD Digital - SVOD Digital - Social Media Physical sales (DVD, Blue-ray, …) Complete Audiovisual content : Films / Series TV magazines/TV news/weather/ sports event Soap Opera Music clips Short forms / User-generated content Higher budget works Lower budget works Legend: expected impact of Gen AI works on the market segment (market penetration of Gen AI works) Note: (1) Audiovisual market excluding audio podcast and video games Flow programmes* Stock programmes* Animated works/Kids shows * Stock programs can be rebroadcast (e.g., fiction, documentaries), while flow programs are usually aired once (e.g., news, sports, games) Advertising 1 Market size Appendix – Detailed methodology and assumptions Audiovisual

96. 96 Study on the economic impact of Generative AI in the Music and Audiovisual industries For each of these segments, a Gen AI penetration rate has been estimated based on use cases expected impact (from low to very high) – Focus on complete AV outputs Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews 2023 market size and 2028 forecasts on segments impacted by Gen AI Fully Gen AI artworks Penetration rate in 2028 2028 Fully Gen AI AV works’ market size 1 Market size x = ‘23 market size ’28 forecasts c. €48Bn Total Gen AI market size for complete audiovisual productions Disclaimer: Market size including royalty-free audiovisual content (mainly videos) Weighed penetration rate, including(1): TV broadcasters €327Bn €298Bn • Flow programmes of which news, weather, feature stories…) • Flow programmes of which sports events, games… • Ads & clips (lower budget works) • Stock programmes – of which films, series… • Stock programmes – of which kids animated works… 4% €11Bn Digital - SVOD €107Bn €161Bn 2% €3Bn • Higher budget AV works (films for cinema…) • Lower budget complete AV works (daily soap opera, reality TV…) • Animated works/Kids' content Digital - AVOD €37Bn €60Bn 8% €4.5Bn • Music/covers videos • Educational/tutorial • Gaming videos • Corporate videos • Others (incl. vlog, comedy) Digital - Social Media €186Bn €230Bn 13% €30Bn • Short form videos (Reels, TikTok) • Long-format videos Appendix – Detailed methodology and assumptions Audiovisual

97. 97 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual collections segmentation and identification of segments on which AI impact is marginal Revenue cannibalisation for creators has been calculated based on a segmentation of Audiovisual global collections Source: PMP Strategy market model, Experts’ interviews 1. Isolation of collections for translators/adapters and screenwriters, and calculation of the revenue loss of these categories based on cannibalisation rates assumptions for each AV works represented 2. For the other works (complete audiovisual outputs): Bottom up & Top-down approach to calculate the potential revenue loss for complete audiovisual outputs : – Breakdown of CISAC 2023 global audiovisual collections by categories and subsegments – 2028 collections forecast for each subsegment based on historical growth rates and future market trends – Estimation of Gen AI outputs’ works cannibalisation rate in 2028 based on use cases and market estimates conducted as part of question 1. Key methodology insights CISAC global collections (for directors & other co-authors) Collections segmentation by use • Exclusion (or marginal impact on the total) from the calculation of the collections’ subsegments where the AI generated works are unlikely to be cannibalized by Gen AI (blurred boxes): • E.g., (i) Marginal impact on radio collections as few collections from radio components likely to be impacted by AI (audio podcasts) (ii) Private copying works may shrink due to shrinking of global revenues but collections will still be from Rationale human creations Private Copying Digital SVOD AVOD Social Media TV & Radio Radio TV Live & Background Background Collections translators/adapters Others Educational use Mechanical reproduction Rental, Public Lending Others Collections screenwriters Revenue loss 2 Appendix – Detailed methodology and assumptions Audiovisual

98. 98 Study on the economic impact of Generative AI in the Music and Audiovisual industries CISAC audiovisual’s collections evolution by categories/revenue streams in current market evolutions | €m, 2018-2028 Based on historical growth and future market trends, CISAC collections for Audiovisual are expected to reach 814m by 2028, with a 24-28 CAGR of c.3% Source: PMP Strategy market model, Experts’ interviews Forecasts 2% 0% 2% 45% 4% 2018 2019 2020 2021 2022 35% 6% 6% 2% 2% 0% 45% 4% 2023 2024 2025 2026 2027 31% 6% 10% 2% 38% 1% 7% 2% 45% 4% 1% 605 597 626 608 646 2028 710 732 757 784 814 690 TV & Radio Private Copying Digital Live & Background Educational use Others For Screewriters For Translators & adapters YoY growth (%) -1% +5% -3% +6% +7% +3% +3% +3% +4% +4% CAGR 18-22 22-24 +2% +5% 0% +3% -1% +2% 24-28 +3% +1% +2% +51% +17% +15% -0 +12% +4% +3% +3% +6% -1% +2% +2% Revenue loss 2 +2% +5% +3% +2% +5% +3% For Directors & other authors contributing to the making of audiovisual works Appendix – Detailed methodology and assumptions Audiovisual

99. 99 Study on the economic impact of Generative AI in the Music and Audiovisual industries The revenue loss for directors and other authors contributing to the making of the AV works has been estimated by applying 2028 estimated cannibalisation rates Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews x = ‘23 collections ’28 forecasts Revenue sub-streams (1), 2023 collections and 2028 forecasts Gen AI works 2028 cannibalisation rates 2028 revenue loss for creators €240m €252m c. €19m(1) (5%) Total revenue loss for directors & other authors contributing to the making of audiovisual works TV & Radio • Flow programmes of which news, weather, feature stories… • Flow programmes of which sports events, games… • Ads & clips (lower budget) • Stock programmes – of which films, series… • Stock programmes – of which kids animated works 4% €9m Digital €41m €82m • AVOD: (weighed penetration rate from market) • SVOD & digital TV • Social media €4m €12m €15m Live & Background • Background audiovisual works (considered as highly at risk as Gen AI will enable to broadcast unlimited content in public places, stores…) 30% €5m Note: (1) Includes cannibalisation for educational use (potential replacement for tutorials, webinars…) resulting in a €1m revenue loss Weighed penetration rate, including(1): 2 Revenue loss €59m €66m Other substreams Marginal to no impact Total collections €352m €415m 5% Appendix – Detailed methodology and assumptions Audiovisual

100. 100 Study on the economic impact of Generative AI in the Music and Audiovisual industries The revenue loss for translators/adapters and screenwriters have been estimated based on a forecast of 2028 revenues and the application of cannibalisation rates Source: PMP Strategy market model, financial reports, Specialized press, Experts’ interviews x = ‘23 collections ’28 forecasts c. €13m Translation / Adaptation €28m €33m • cannibalisation rate equivalent to the weighed penetration rates from market size calculations 41% €13m Screenwriting €311m €366m • cannibalisation rate equivalent to the weighed penetration rates from market size calculations 19% €70m Total revenue loss for translators/adapters Total revenue loss for screenwriters c. €70m ’23 revenue sub-streams and ’28 forecasts Gen AI works 2028 cannibalisation rates 2028 revenue loss for creators c. €0.1 Bn (12%) Total revenue loss in 2028 for audiovisual creators 2 Revenue loss Total collections (incl. the three categories) €690m €814m Appendix – Detailed methodology and assumptions Audiovisual

101. 101 Study on the economic impact of Generative AI in the Music and Audiovisual industries Additional Revenue loss calculation for screenwriters, directors & other co-authors Source: PMP Strategy analysis, FERA, SAA, European Audiovisual Observatory, Experts’ interviews ‘23 AV production market size (budget) and ‘28 forecasts Gen AI ’28 cannib. Rates by segments ’28 creators’ revenue loss (and %) Share of main segments going to AV creators x x = €209Bn €229Bn ’23 market size ’28 forecasts Includes streamers & broadcasters original content spending, public fundings and fiscal incitation (i.e., includes also independent production) For directors & other co-authors c.4-5% For screenwriters Includes upfront payment and contingent payment from producers c.4-5% For directors & other co-authors For screenwriters • Unscripted works • Higher production budget • Lower production budget • Animated works • Higher production budget • Lower production budget • Animated works €1.9Bn €1.2Bn B Revenue loss 2 15% 20% Appendix – Detailed methodology and assumptions Audiovisual

102. 102 Study on the economic impact of Generative AI in the Music and Audiovisual industries Additional Revenue loss calculation for authors of audiovisual translations, adaptations and subtitles Source: PMP Strategy analysis Gen AI ’28 cannib. Rates by segments ’28 creators’ revenue loss (and %) x = ’23 worldwide (non CMO-collected) revenues and ‘28 forecasts €1.3Bn • Higher production budget €1.9Bn €2.2Bn • Lower production budget ’23 market size ’28 forecasts(1) Revenue loss 2 56% Appendix – Detailed methodology and assumptions Audiovisual

103. 103 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual Gen AI providers’ revenues – 2023-2028, €Bn Calculation methodology for the audiovisual segments (1/2) Complete video/audiovisual works Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights Runway 2028 active users' estimation Runway 2028 ARPU Runway market share in prompt-to-video market c. €3.5Bn 2028 associated players’ total revenues c.4m - 30% 40€ 15% x / = Video/Audiovisual software market size 2023 2028 €1.4Bn €1.9Bn 2028 Generative AI tools penetration rate €0.8Bn 2028 associated players = total revenues 40%(1) / Mid x Note: (1) Considered as slightly lower as in audiovisual, due weaker maturity of tools to date, and an estimated even more expensive price for these tools in 2028 1. Mass public tools (also suitable/used for professional purposes) 2. Gen AI tools/software as an assistant (support in the creative process) Gen AI services 3 revenues Appendix – Detailed methodology and assumptions Audiovisual

104. 104 Study on the economic impact of Generative AI in the Music and Audiovisual industries Audiovisual Gen AI providers’ revenues – 2023-2028, €Bn Calculation methodology for the audiovisual segments (2/2) Dubbing/Subtitling and Screenwriting Source: PMP Strategy market model, Experts’ interviews, Specialized press, Business Research Insights Screenwriting software market size 2023 2028 2028 Generative AI tools penetration rate 2028 associated players = total revenues €140m €230m 60% / High €125m x Dubbing/Subtitling software market size (incl. social media) 2028 Generative AI tools penetration rate(1) 2028 associated players = total revenues x €1.3Bn €1.9Bn 41% €0.9Bn 2023 2028 Dubbing / Subtitling tools Screenwriting tools Gen AI services 3 revenues Note: (1) Based on weighed penetration rate calculated in the market segment | (2) Higher rate as in the market calculation, as such software are widely used for audiovisual content that is not intended to generate revenue or is not widely distributed (corporate content, content for social networks) Appendix – Detailed methodology and assumptions Audiovisual

105. 105 Study on the economic impact of Generative AI in the Music and Audiovisual industries Detailed methodology and assumptions Glossary Detailed list of interviews conducted PMP Strategy presentation Appendix

106. 106 Study on the economic impact of Generative AI in the Music and Audiovisual industries Glossary | Main abbreviations and definitions (1/2) • AV: Audiovisual • ARPU: Average Revenue Per User • B2B (Business-to-Business): Refers to transactions between businesses, such as a manufacturer selling to a wholesaler. Examples include companies providing office supplies to other businesses • B2C (Business-to-Consumer): Businesses selling products or services directly to individual consumers. Examples include online retailers like Amazon. • Buy-out: A one-time payment for the full rights to use a creative work, with no future royalties owed to the creator. • CAGR: Compound Annual Growth Rate • CMO: Collective Management Organization • Deep Learning: A subset of machine learning involving neural networks with many layers, enabling the analysis and learning from large amounts of complex data. • DSPs (Digital Service Providers): In the context of music streaming, Digital Service Providers are online platforms that distribute and stream music to listeners. Examples include Spotify, Apple Music, and Amazon Music. • GAFAM: Acronym for Google, Apple, Facebook, Amazon, and Microsoft • Gen AI (Generative Artificial Intelligence): AI systems that generate new content based on training data. • Input: The data or information fed into an AI system or algorithm for processing and analysis. • LLM (Large Language Model): A type of artificial intelligence model trained on vast amounts of text data to understand and generate human language. • Machine Learning: A branch of artificial intelligence where algorithms learn from and make predictions or decisions based on data. • NLP (Natural Language Processing): A field of artificial intelligence focused on the interaction between computers and humans through natural language. • OTT: Over the top, self-distribution model outside the operator set-top box: content accessible directly through an app/website on all devices (smart TVs, smartphones, tablets, etc.) Appendix - Glossary

107. 107 Study on the economic impact of Generative AI in the Music and Audiovisual industries • Output: The result or product generated by an AI system, such as text, images, or other data. • Pay-per-view (TVoD): a television service in which viewers are required to pay a fee in order to watch a specific programme. • UGC (User-Generated Content): Content created and published by users rather than by professional creators or brands, often shared on social media and other online platforms. • VOD: Video-on-demand – AVOD: Advertising video on demand, i.e. advertised-funded digital video platforms (YouTube, Social Media) – BVOD: Broadcaster Video On Demand, free-access streaming platforms from local broadcasters (VRT MAX, VTM Go, Go Play) – HVOD: Hybrid video on demand, combining several business models (advertised-funded and subscription/consumer-funded for instance) – SVOD: Subscription video on demand, traditional streaming platforms, including international players (Netflix, Disney+, etc.) and local players (Streamz) – FAST: Free ad-supported streaming TV Glossary | Main abbreviations and definitions (2/2) Appendix - Glossary

108. 108 Study on the economic impact of Generative AI in the Music and Audiovisual industries Detailed methodology and assumptions Glossary Detailed list of interviews conducted PMP Strategy presentation Appendix

109. 109 Study on the economic impact of Generative AI in the Music and Audiovisual industries Interviews | CMOs (1/2) Source: Experts’ interviews & workshops Name Company Repertoire Marie-Anne Ferry-Fall & Thierry Maillard ADAGP Visual Arts Dean Ormston & Richard Mallett APRA AMCO Music Christian Zimmermann & Reema Selhi DACS Visual Arts Ricardo Gómez Cabaleiro DAMA Audiovisual Tobias Holzmüller & Kai Welp GEMA Music Kazumasa Izawa & Kay Yamaguchi JASRAC Music Chu Ga Yeoul & Seon Cheol Hwang KOMCA Music Andrea Czapary Martin & John Mottram PRS Music Alexandra Cardona Restrepo REDES Audiovisual Géraldine Loulergue-Husson & Patrick Raude & Sandrine Sandoval SACD Audiovisual Héloïse Fontanel Sacem Music Julien Dumon Sacem Music David El Sayegh Sacem Music Julien Lefebvre Sacem Music Appendix – Interviews

110. 110 Study on the economic impact of Generative AI in the Music and Audiovisual industries Source: Experts’ interviews & workshops Name Company Repertoire Annabell Lebethe SAMRO Music Cristina Perpiñá-Robert Navarro SGAE Transversal Matteo Fedeli & Fabrizio Zavagli & Adriana Galli & Andrea Marzulli SIAE Transversal Jennifer Brown SOCAN Music Jürg Ruchti SSA Audiovisual Marcelo Bastos Castello Branco & collaborators UBC Music Sylwia Biadun ZAPA Audiovisual Vianney Beaudeu & Raphaël Léaupard LaScam Audiovisual Maria Garateche Argentores Audiovisual Richard Combes ALCS Audiovisual Interviews | CMOs (2/2) Appendix – Interviews

111. 111 Study on the economic impact of Generative AI in the Music and Audiovisual industries Interviews | Tech players Appendix – Interviews Source: Experts’ interviews Name Company Repertoire Àlex Loscos BMAT Music Ed Newton Rex Fairly Trained Transversal Ryan Groves Infinite Album Music / Audiovisual Nathalie Birocheau Ircam Amplify Music Alexandre Défossez Kyutai Transversal Philippe Guillaud Matchtune Music Eric Samson Microsoft Transversal Christophe Müller & Kevin Montler YouTube / Google Transversal Appendix – Interviews

112. 112 Study on the economic impact of Generative AI in the Music and Audiovisual industries Interviews | Production, Distribution, Publishing companies Source: Experts’ interviews Name Company Organisation type Pierre-Michel Levallois BAM Music Production Company Aurélien Hérault Deezer DSP Mathieu Taieb Dubbing Brothers Production Company Perrine Guyomard Ex-Warner / Sacem Lab Production Company / CMO Tiphaine Des Déserts Getty Image Production & Publishing Michael Turbot Sony Computer Science Laboratories Production Company Anne Jouanneau Sony Music Publishing Publishing Company Appendix – Interviews

113. 113 Study on the economic impact of Generative AI in the Music and Audiovisual industries Interviews | Institutions, Legal bodies and other organisations Source: Experts’ interviews Name Company Repertoire Marion Carré Ask Mona / Commission Européenne Transversal Sylvie Fodor CEPIC Visual Arts Cécile Lacoue CNC Audiovisual Arshia Cont Ex-Ircam / Antescofo Music John Phelan ICMP Music Lauri Rechard, Abbas Lightwalla IFPI Music Alfons Karabuda NIM / ECSA Music Alexandra Bensamoun Paris Saclay / Commission interministérielle de l'IA Transversal Benoît Carré SGYGGE / Ministère de la Culture Music Isabelle Wekstein-Steg WAN AVOCATS Transversal Juliette Prissard Eurocinema Audiovisual Céline Despringre SAA Audiovisual Pauline Durand-Vialle FERA Audiovisual Gilles Fontaine European Audiovisual Observatory Audiovisual Eduardo Senna & Matheus Leopardi Senna Advogados Audiovisual Appendix – Interviews

114. 114 Study on the economic impact of Generative AI in the Music and Audiovisual industries Detailed methodology and assumptions Glossary Detailed list of interviews conducted PMP Strategy presentation Appendix

115. 115 Study on the economic impact of Generative AI in the Music and Audiovisual industries Your business environment is changing faster and faster. We already know that we will not work or consume tomorrow as we do today. We have learned that we need to be able to adapt quickly to major disruptions, which are unpredictable by nature. And that it is no longer acceptable to impact business performance without taking into consideration the world we live in. At PMP Strategy, we are committed to working with senior executives to achieve both goals and create a positive impact. Just as you are, we are true experts in your market. We apply rigorous analysis, leveraging our high level of competence and understanding. We believe that there is no positive impact without great conviction and the total mobilisation of a diversified team. We cannot address today's problems with yesterday's answers. We are dedicated to finding tailor-made and innovative solutions, as if we were doing it for ourselves. We are always focused on working with you and your teams, hand in hand, with the entrepreneurial spirit that motivates us. We strongly believe in diversity, human commitment and openness. These are the values that forge the strong relationships and mutual trust that we cultivate with our clients. PMP Strategy is a renowned consulting firm in the Media & Cultural industries, helping its clients adapt their business models and organisations in a market disrupted by digital "pure players". A Strategic Consulting Firm Telecom Media Tech Transportation & Mobility Energy & Industry Decarbonization Financial Services & Institutions Strategy & Transformation Private Equity Digital, Data & Customer Experience CFO Advisory and Integrated Performance CSR-ESG Industries Capabilities PMP Strategy: An International Player Europe Africa & Middle East North America Dubai, Casablanca New York, Montreal, Seattle, Toronto Key figures 11 Offices. +150 Consultants. +100 Expert Advisors. Paris, London, Madrid, Brussels, Luxembourg Cultural & Creative Industries

116. 116 Study on the economic impact of Generative AI in the Music and Audiovisual industries Philippe Curt pcurt@pmpstrategy.com Helene Moin hmoin@pmpstrategy.com François Cousi fcousi@pmpstrategy.com Raphaël Flabeau rflabeau@pmpstrategy.com CISAC Communications +33 1 55 62 08 50 communications@cisac.org Contacts