Summary of discussions

Working Group 1 (WG1) analysed the role of intermediaries (e.g. streaming and Web3 platforms) and Artificial Intelligence (AI) systems in the value chain, both at the consumption stage (e.g. algorithmic recommendations), the production stage (e.g. analysis of usage data) and the creation stage (e.g. image and text generators). In other words, the aim was to examine the changes brought about by digital society firstly on creation itself, and secondly to examine the impact on the value chain allowing artistic creation to have an economic existence.

Two themes were examined: the role of platforms in funding of creation, and the role of recommendation algorithms in the visibility of works, the question of AI having been left aside for because of time constraints.

The role of platforms in financing creativity

The digital revolution has profoundly altered the relationship between artists and their audiences, as well as the role of intermediaries. The horizontal communication made possible by the Internet calls into question the dominant “romanticised” monetization model of the 20th century, according to which the artist creates and leaves it to others to monetise his or her talent (e.g. publishers or major labels). Today, the artist has become an entrepreneur who must interact effectively with the distribution bodies on the one hand and his or her audience on the other. The economic model, whereby artists had to be represented by a publisher or a major label to gain access to the market, is giving way to a new structure in which artists have more means at their disposal to become autonomous. Moreover, self-production is growing strongly at the same time as traditional production, leading to an overproduction of highly homogenous content. In the book industry, for example, the case of Amazon is telling. Publishers are sometimes reluctant to charge authors to publish their work, particularly in the academic world. Sites such as lulu.com give authors a great deal of leeway in terms of how their work is transmitted, while at the same time enabling them to reach their target audience via e-commerce ³ .

The digital cultural economy is also characterised by a certain rematerialisation (e.g. in the form of tangible objects such as records or printed matter, or live performances). Often, an artist publishes a work "free of charge" (e.g. under a creative commons licence ⁴ ). A platform then ensures its promotion and visibility. Finally, the work is "materialised" when the artist is hired to perform in public, and the artist returns to the physical world to monetise their work via a performance fee. No artist can afford to stop "touring". During live performances, the artist creates value through a certain rarity and authenticity of experience that is not available online.

These changes call into question the traditional role of the publisher. Artists now have ways of bypassing the publisher to reach their audience. The role of platforms tends to be limited to that of a gatekeeper. Their value now lies in their social value. Value is in the selection process by which the publisher chooses one work over another. This leads other market players to trust in and pay for the work. In other words, what has value is is the reputation of the intermediary who associates his credibility with the work he is going to edit and then publish. This is not an objective value, but the social value of the intermediary.

The role of recommendation algorithms in access to creation

Platforms can represent a threat to cultural diversity, especially when they use recommendation algorithms that do not have parameters to ensure a diverse content. Algorithms themselves are not a threat, but they must be used wisely. The opacity of the algorithm, or more generally the lack of transparency about its function, is a limiting factor to guarantying cultural diversity. On a technical level, the algorithm often relies on the following theoretical approaches: (i) Use of metadata: this involves examining the content that has been liked or consulted, often by attaching keywords to it, and then subsequently suggesting content to which identical or equivalent keywords have been assigned; (ii) Content analysis: In the case of music, for example, this involves calculating the number of beats per minute, the instruments used and other elements in order to suggest something similar; (iii) Profiling: if a user listens to a track belonging to a typical profile, there is a good chance that they will conform to it. (iv) The algorithm searches the raw signal of the content consulted for similarities with other content (sometimes on the basis of an image, a sound or their succession) and suggests, without the human really knowing what it has based itself on, content that the machine deems similar.

Recommendation algorithms also tend to isolate users in bubbles, in which they always consume the same types of content because it is suggested to them. Algorithms also influence the artist, who bends to their constraints to gain visibility. The issue is therefore not an individual one, but one of cultural policy. In this context, several initiatives aimed at protecting cultural diversity can be mentioned. In France, the “pass culture” offers young people aged 15-18 an app through which they can access a certain amount of money to buy cultural services offered based on human and algorithmic recommendations aimed at diversification ⁵ . Several influencers and the playlist system are also a part of this movement.

Recommendation algorithms are also changing the way artists create. The problem for artists is no longer finding the right door to knock on to reach their audience and make a living from their work, but acquiring visibility in a context where communication is horizontal, global and immediate. However, the visibility of content depends on algorithms that are generally opaque. In such conditions, artists are unable to determine how they can gain this visibility. They may even be tempted to produce less original, more calibrated content to please the algorithm. With a plethora of immediate content on offer, the public tends to rely on black boxes that do not necessarily aim to diversify or satisfy their expectations, but to maximise the user's advertising. It should be noted that the promotion of certain content linked to user preferences can take different forms: in the pre-digital era, programmers relied on radio and TV programme schedules. In the digital age, intermediaries offer content using algorithms, or even AIs, that automate recommendations to users ⁶ . However, the vast majority of algorithms are designed to maximise the user's exposure to commercialisation, whether this is a subscription or exposure to advertising, since they generally come directly from the entity offering the content. One way forward would be to break this concentration.

Recommendations

One public intervention strategy would be to separate the providers who supply content (e.g. artists, publishers, producers) and those who distribute it (e.g. social and streaming platforms). In other words, we need to separate recommendation from distribution and give users and creators the opportunity to decide on the methods of recommendation to which they are exposed. Users connecting to a distribution service would have a choice of third-party recommendation tools (recommender) with multiple objectives and functions.

Legally, this obligation to dissociate recommendation and dissemination would stem from the fundamental right to digital self-determination, i.e. the ability of citizens to decide how they access information. The authorities would be responsible for implemented by interpreting the existing legal framework (e.g. competition law) or specific laws to be adopted (e.g. EU Digital Services Act (DSA), which enshrines a new right to parameterisation).

However, such a right runs the risk of being ineffective. Just think of the experience of cookies consent banners that Internet users do not read. So, rather than a right, it would be better to establish an obligation of parameterisation, whereby the entity providing a service must, following the example of what Microsoft had to do with its Internet Explorer browser, offer a choice of recommenders to the user who uses its services. The transparency of the algorithm would nonetheless remain a problem. Few users would have the confidence to make their own choices, even if they were given the means. It would therefore be interesting to consider a system, similar to the systems proposed by guarantee marks, that would enable the certification of recommendation offers. The proposed system would have the goal of preventing the entities that recommend content from being the ones that pay the search engines. These rules could finally be given concrete form by means of standardisation or even certification procedures. Standardisation or certification could be promoted and even controlled by UNESCO, which organises aspects of access to culture, and the various institutions that safeguard fundamental rights, such as the Council of Europe or the ECHR ⁷ .

Summary of discussions

Working group 2 (WG2) focused on the role of the public in the governance and funding of culture. The participants took as their starting point the observation that digital technologies and web tools (e.g. streaming platforms, Web3, platforms, curation and algorithmic recommendation, generative artificial intelligence) are shaking up practices and putting into question the place of the audience within the cultural and creative sector. We need only think of fan fiction, where a community of fans extend the work of an artist, and which has developed thanks to participative platforms. At the same time, so-called "generative" artificial intelligence (or GenAI) generate new cultural content with user data. These innovations are decisively transforming the relationship between audience and artist. The issues at stake include the forms of artistic creation, the status of the artist, remuneration models and our definition of creativity. Should users, for example, be paid for their contribution to the databases on which artificial intelligence (AI) is based? Based on these observations, WG2 focused on three main areas: artistic creation, the circulation of works and the remuneration of the various players.

The debates were organised around three main points: audience creation, developments in copyright and the limits of scalability.

How are audiences formed?

The working group's discussions first emphasised that audiences are not given as is, but rather are formed through the intervention of the different entities that shape them. Moreover, there is not one audience on the Internet or elsewhere, but multiple audiences. Every artist should have the means to compose his or her audience. This first statement requires two elements.

Firstly, this presupposes that artists have access to the data collected from the consumption and circulation of their works. This data is essential for guiding artists' development strategies and nurturing links with their audiences. Secondly, platforms and other web intermediaries should be more transparent about how their audiences are generated. This would give artists a better understanding of the logic behind the circulation of their work.

The working group also noted that audience data is tending to become the basis for calculating the remuneration of the various players in the cultural markets. Not does data make it possible to understand online cultural consumption practices, but they are also used to decide how income from the sale of advertising or subscriptions should be distributed. There is therefore a direct link between the way in which audiences are shaped and the remuneration of works. From this point of view, remuneration models could also be reviewed, as illustrated by the current debates on music streaming around the introduction of new, fairer remuneration models for artists ⁸ . Studies show that it is mainly intermediate artists who have the most difficulty in being remunerated.

The group's discussions also highlighted the fact that cultural actors are often in a weak position vis-à-vis web platforms. One strategy for public intervention could be to strengthen the cultural actors vis-à-vis the platforms to redress the balance of power. Several avenues were raised during our discussions. Recent European legislation (the General Data Protection Regulation (GDPR) and the Digital Markets Act (DMA)) is also moving in this direction ⁹ . It could go further, for example, with the introduction of a tax on the production of generative AI to fund artistic creation. Finally, the public production of shows is heavily regulated (number of persosns, etc.). The same could apply to the activity of platforms. This type of regulation already exists for radio, for example (anti-payola law, national artist quotas, etc.).

Who should be protected by copyright?

The discussions touched at length on the definition of copyright. Some of the issues relating to audiences concern the way in which the notion of the author is redefined by new participation practices. Three questions were raised in particular:

1. Should copyright be extended to all forms of production and co-production? For example, should all those involved in the creation of a fan fiction be covered by copyright?

2. Should copyright apply only to the work of professional artists or also to that of amateurs?

3. What about content produced by AI? Should it too be protected by some form of copyright?

The various discussions highlighted the fact that the boundaries between professional and amateur practices are becoming increasingly blurred. The advantage is that copyright does not operate on this boundary, but on the idea of creation. For some members of the group, it is important to encourage access to copyright (including through creative commons licences).

Regarding generative AI creations, group members identified several issues. Some members of the group believe that, if generative AI creations are not protected by copyright, they will have a competitive advantage over "human" creations. Since no rights will have to be paid, they will be cheaper. Other members of the group stressed that it is difficult to distinguish between these different forms of cultural content. Where does human production end and machine production begin? Can writing prompts be considered a creative activity? Doesn't every creative activity benefit from the intervention of a machine (instrument, computer assistance, etc.)?

The group noted that these issues may not concern the most innovative and disruptive content, but more common forms of cultural content (background music, book covers, poster design, etc.). However, there should be no difference in the treatment of works from this point of view, especially as these are in fact major parts of the creative industries that are directly affected by these issues.

Promoting communities in the face of the scalability of artwork

The members of the group pointed out that the organisation of most of the major digital platforms is based on the idea of scalability. Faced with this model, some players are proposing alternatives whose economic model is based on smaller communities. These make it possible to create more direct links (a form of "AtoA" model, amateurs to artists). The working group noted that scaling up is not necessarily a foregone conclusion. New ways of funding creation are being developed and could provide solutions in the future. One public intervention strategy could be to support these initiatives directly. This model may also have the advantage of responding better to the challenges of the environmental crisis. In addition, current approaches favour individual remuneration and make it very difficult to work together. Models based on cooperation seem to be an avenue worth exploring. These schemes also make it possible to envisage other forms of artistic commissioning that could be promoted by publicly-funded cultural institutions, as is the case with the Festival Réel in Villeurbanne (France), the Société des Nouveaux Commanditaires or decentralised autonomous organisations (DAOs) ¹⁰ . It would also be possible to imagine the development of platforms supported directly by public funds.

Recommendations

The working group has drawn up three scenarios with recommendations. These relate to promoting human creativity on the Web, improving transparency and strengthening the capacity for public action.

1. Promoting human creativity on the Web

A first area of intervention could be aimed at promoting human creation on the Web. The main aim would be to protect and promote human practices in the face of AI-generated content. On the one hand, the aim is to avoid unfair competition from AI-generated content for artists in terms of cost. On the other hand, the challenge is to encourage aesthetic experiences and creative practices among the general public. The main obstacle to this is the difficulty of distinguishing between what is machine intervention and what is human creativity.

One idea for implementation that emerged from the discussions was the introduction of a tax on generative AI that would help to fund creation (e.g. in the form of an extension of copy levies for AI-generated content).

2. Promoting the transparency of digital platforms

The second area for action proposed by this working group is to promote transparency in cultural production. It is based on three observations. Firstly, more and more content is being created by AI without this being clearly established from the point of view of the public. Secondly, there is a need to improve the traceability of cultural content and its circulation. Thirdly, there is a need to raise public awareness of these issues. The group identified two main difficulties in this area. On the one hand, the task is significant and it is difficult to define exactly how far transparency efforts should extend. Secondly, it is impossible to make a complete distinction between human and non-human interventions.

The working group is proposing three avenues to explore in this area:

1. The creation of labels to identify types of cultural content, particularly AI-generated works.

2. The obligation to declare the nature of works when registering with a copyright collective management organisation.

3. Strengthening consumer rights (consumer law) and, in particular, content creators' duty of authenticity to the public.

3. Strengthening the audience’s ability to take action

The third area for action is also based on three observations. Firstly, audiences are shaped by different actors, and there is a need for a better understanding of the role of each audience in these processes. Secondly, the promotion of works, diversity and discoverability are central issues for digital platforms. Thirdly, data is an issue not only for understanding audiences, but also for remunerating the various players. The main challenge in this area is data protection and privacy.

The working group came up with four implementation ideas:

1. Strengthen the portability of user data to encourage circulation between platforms.

2. Demand greater transparency of recommendation algorithms and develop a right for users to configure them.

3. Encourage access to data for cultural players and offer greater transparency on remuneration, including for the general public.

4. Better supervision of content promotion and avoidance of payola logic.

In conclusion, the group's discussions highlighted the importance of audiences in the funding and governance of the worlds of culture. While much of the current debate focuses on the creation and circulation of works, audiences appear to be an important lever for shaping the future of culture and new forms of public intervention.

Summary of discussions

The main task of Working Group 4 (WG4) was to analyse the role of copyright and related rights management companies in the value chain, in particular whether and how to adapt this role in light of new modes of creation and distribution (e.g. streaming, artistic and participatory NFTs). The aim was to analyse the process and criteria used by collective societies to negotiate remuneration with operators, exhibitors, platforms and distributors, and to identify the relevant criteria for distributing royalties within these collective societies, taking into account current practices or those to be envisaged. The aim was to come up with innovative solutions that would benefit artists and creativity, more specifically through policies aimed at the authorities and professionals. In particular, this meant analysing the new models of data trusts aimed at pooling certain data.

Once these contours had been determined, the group was able to consider its strategic direction, i.e. whether it should focus on protecting artists against market fluctuations, and whether Web3 technology could serve as a protective mechanism for artists made vulnerable to exploitation by large institutional entities. These questions led them to consider limiting or even eliminating the intervention of intermediaries who deal primarily with objects, such as art galleries, or individuals, such as labels. The working group noted that only a limited number of artists are really represented and/or protected by galleries or labels, even though the latter often have the influence and networks needed to disseminate an artist's work. In reality, the promotion and sale of artistic works seems to be closely linked to entrepreneurial artistic practices specialising in the creation of branded content, which is then distributed through renowned market channels such as art festivals. In this context, associations of independent artists, equipped with legal mechanisms and blockchain technology designed to protect their own interests as well as those of their peers, could be considered as a relevant alternative. For WG3, the sharing model observed on platforms such as Soundcloud is also interesting in this respect. These platforms represent small communities where users publish their work and earn points by sharing other members' tracks. As a result, they can use these points to promote their own creations, creating a cycle of collaboration and support. However, this raises questions about the ability of blockchain and Web3 technology to provide an appropriate space for implementing such a system on a wider scale. At the same time, it should be noted that emerging artists are actively seeking sustainability, but face a market that appears complex and confusing to new participants. In this environment, understanding contractual agreements or knowledge of blockchain technology requires specific knowledge. In this sense, the members of the working group observed that non-fungible tokens (NFTs), which could offer opportunities to artists, remain largely underused due to their inaccessibility and the difficulties associated with understanding how they work. What's more, these opportunities are often not financially viable. This forces artists to live with in precarious financial situation, the stability of which varies according to sales, reinforcing the need for a sustainable alternative.

Recommendations

The avenues explored by the working group led to two proposals designed to provide a partial response to the current problems of artistic remuneration mentioned above: a redistributive payment system called of "F.A.I.R" (Federated Art International Redistribution) and the creation of autonomous and temporary art zones. More specifically, these solutions function as artist-oriented responses to the current art marketing system using Web3 technologies. These proposals call for further development. They are the result of a discussion enabled by the WKD Congress and mark the beginning of a broader discussion aimed at strengthening the sustainability and prosperity of artists in the near future.

1. Redistributive payment system

The working group's initial concept focuses on the limitations of the current gallery-based art economy, while offering an alternative to emerging artists who face obstacles in establishing themselves in the market and generating income from the sale of their work. To remedy this, the group is proposing a redistributive payment system operating under the name "F.A.I.R" (Federated Art International Redistribution). This service aims to standardise the redistribution of funds to artists from customers and other market players using smart contracts on the blockchain. Specifically, this system can split the initial payment into smaller payments, allocating them to various beneficiaries, including the artist, developers, charities or other designated funds.

In other words, part of the revenue from the sale of a work of art would serve as a source of capital that the system would distribute according to predefined conditions. The use of smart contracts means that complex conditions can be specified, and redistribution tailored to the specific needs of artists, such as co-ownership of a work or royalties on future sales (comparable to droits de suite). In addition, the division process is designed to be fully automated, eliminating the need for intermediaries, reducing transaction costs and processing times, all of which act as incentives for the adoption of this service. Smart contracts also contribute to greater transparency, in that key aspects of fund operations are publicly accessible and verifiable on the blockchain. However, the disadvantages or risks associated with this system include the possibility of monopolising artists' payments. In addition, it is imperative to consider the limitations of smart contracts, such as the time required for programming, which can increase costs for users. In addition, the legal applicability of the contract in different jurisdictions needs to be carefully considered. Similarly, the choice of the most suitable blockchain for implementing such a system also requires careful evaluation.

However, we can see promising applications of this mechanism for redistributing funds to remedy gaps in artists' access to social benefits. By way of illustration, this framework could facilitate the formation of a recurring monthly income, financed through smart contracts associated with transactions, thus offering emerging artists a form of universal basic income and older artists a retirement pension. While technically feasible, it should be borne in mind that this functionality raises complex legal challenges and governance issues. Furthermore, the working group recognises that for this idea to be implemented and embraced by artists, the proposed system needs to be integrated into a broader arts organisational framework and become operated by entities, such as the Mozilla Foundation or Pro Helvetia ¹¹ . Indeed, such entities would lend credibility to banking institutions and legal authorities, and generate confidence among the target audience while simultaneously establishing links with a vast network of artists. Furthermore, it is imperative that the project is both accessible and understandable to achieve the desired effectiveness; the working group observed that many Web3 and Blockchain initiatives are technically sound, but fail to convince artists to integrate them into their ecosystem. The success of such a venture may therefore also depend on design thinking and a user experience tailored to meet the needs of artists who may not be familiar with these technologies.

2. Temporary and autonomous art zones

The second proposal is a complementary approach to the previous one and stems from the power dynamics of the art market. Indeed, an attempt to establish an alternative art trading system that competes directly with the existing global infrastructure is likely to fail due to the inability to achieve the critical mass necessary to generate a network effect. However, this second proposal does not seek to compete with the global system, but instead focuses on smaller, more localised areas to cultivate 'bubbles' called 'T.A.A.Z.' (Temporary Autonomous Art Zones). These 'bubbles' can be defined as small-scale, interconnected ecosystems in which artistic transactions take place through an exchange of tokens uniquely associated with a specific zone, encompassing diverse identities, names and values that reflect the characteristics of a zone. These transactions would take place within a dedicated exchange platform similar to Uniswap ¹² , ensuring a fully decentralised operational framework. In essence, a token is generated through market transactions or when services are rendered to an artist, such as legal advice.

These tokens can then be spent on other works of art, without needing to be converted into traditional currency, because they have an intrinsic value arising from their negotiability within a thriving local art enclave. The aim is to encourage the acquisition of works of art and support their circulation, rather than treating works of art as static assets. The ambition is to create locally interconnected ecosystems that thrive over months or years, facilitated by a token-based structure. While this may indeed stimulate the local scene and potentially generate additional financial opportunities for artists, it should be noted that this approach does not offer monetary compensation as such, but rather creates an environment based on community exchange and mutual aid.

Summary of discussions

The main task of Working Group 4 (WG4) was to analyse the role of copyright and related rights management companies in the value chain, in particular whether and how to adapt this role in light of new modes of creation and distribution (e.g. streaming, artistic and participatory NFTs). The aim was to analyse the process and criteria used by collective societies to negotiate remuneration with operators, exhibitors, platforms and distributors, and to identify the relevant criteria for distributing royalties within these collective societies, taking into account current practices or those to be envisaged. The aim was to come up with innovative solutions that would benefit authors, artists and creativity, more specifically through policies aimed at the authorities and professionals. In particular, this meant analysing the new models of data trusts aimed at pooling certain data.

To begin the discussions, the working group began by mapping collective management in terms of the elements that produce value, that is to say works, metadata, usage data and users' personal data to establish the starting point before embarking on the thought process.

Collective copyright management usually concerns the works themselves. However, collective societies also produce metadata about the works that could be used, controlled or modified as required. The use and availability of this metadata by collective societies could work towards greater transparency with regard to the works used. Metadata monitoring would therefore be relevant. Adding metadata management to the initial management of works dematerialises the practices of collective societies, bringing them into the digital age: indeed, metadata management requires specific tools and technologies.

The group thus identified which elements of value (works, work metadata, usage data, personal data, geographical data) are managed by one or other form of collective societies, i.e. individual management (authors and artists) and collective management (operators). However, this mapping provoked strong reactions within the group due to the lack of distinction between the different art sectors ¹³ , with some participants considering that the sectors are not all affected in the same way ¹⁴ . It was decided to adopt a cross-sectoral approach to the data common to the different sectors and to focus on the reality of business models.

The value of content-related data was born with the development of the Internet and the emergence of social networks. Because of their intrinsic link with the digital world and its development, user data and the personal data of users are major elements in the exploitation models that concern all sectors of the arts. Discussions revealed a certain difficulty in carrying out a general cross-sectional analysis without going into the specific features of collective rights management in different geographical areas ¹⁵ . The group also questioned the identification of the entity with the authority to decide on information sharing, particularly regarding competition law. In this respect, it should be remembered that it is not forbidden to share certain data, and that operators are sometimes involved in collective management from the outset; this is the case, for example, in the music sector (SACEM ¹⁶ ) or the book sector (ProLitteris, CFC ¹⁷ or Reprobel).

The discussion then refocused on the real purpose of collective management, namely maintaining economic and professional bargaining power (strategies), on ways of improving it and on the importance of remunerating authors and artists. Rather than adopting a purely - “contractualist” - vision, it was suggested that artists should be helped by adapting their rights on the four levels of value: works, metadata, usage data and users' personal data. The aim would be to direct innovations towards the collection society management model itself and to ensure that the protection of artists' rights is strengthened on this basis; the idea is to give them access to what is rightfully theirs. Incorporating the monetisation of these additional elements would also enable them to regain power over the music market as a whole. It would also be important to increase the value of content creation in relation to the value of content distribution.

The management of authors and artists' works and services, as well as other elements of value, could make it possible to improve their remuneration and socio-professional situation. Management methods vary from country to country, depending on the legal system or management capacity. In Europe, for example, collective management is particularly well developed. The group therefore considered whether new management models for non-transferable rights such as collective licences ¹⁸ or extended collective agreements ¹⁹ (Directive 790/2019) should be generalised, i.e. alternatives halfway between the law and contractual provisions.

Collective licensing was discussed first. It was made clear from the outset that the territorial aspect is not decisive in the music industry, since online rights are often represented by other companies (linked to the majors) rather than the traditional ones (such as SACEM, GEMA ²⁰ , SUISA ²¹ or others). To obtain a licence, the operator must agree to provide usage data; moreover, there is currently no monetisation of users' personal data. To ensure a fair distribution of value, more attention needs to be paid to the platforms' business models. More specifically, in theory this could involve acquiring data from platforms and then exploit it. The group therefore wondered whether it would be appropriate to include a fee for the use of the data in the collective society fee. However, there are a number of negative aspects to this hypothesis. Firstly, it is not always possible to check how consumers' money is divided between the purchase of products and the purchase of audiovisual services (e.g. the case of Amazon Prime). Secondly, when we talk about "proportional remuneration", the percentage will always depend on what is known as the "calculation basis" for the remuneration. Thirdly, this solution would raise data protection issues, particularly in the event of data being resold. This would mean replicating the platforms' business model.

However, the group wondered why the value derived from the exploitation of data should only benefit the platforms. It was also pointed out that the notion of "content" is not treated as such by social networks, and that it is no longer the most important part of the value creation system (business model). The business model today is to exploit content in the most profitable way. This now means playing by the rules set up by the platforms, and therefore requires personal data to be valued in negotiations. For certain members of the group, to achieve this, we need to create a degree of competitiveness between collective societies and platforms. In this respect, the issue of transparency plays an essential role. One possible solution would be to improve the legal framework by providing for public oversight of platforms and management companies, following the example of the recent amendment of Brazilian legislation on collective societies to work towards greater transparency.

Obtaining data on the use of works is also an important need in the audio-visual field. The chain of responsibility has been put to the test, leading in particular to difficulties in establishing authorship rights over the work with regard to metadata. As far as the book market is concerned, the digitisation process for literature and comics is the slowest in comparison with the aforementioned artistic fields. This is because streaming subscription services for music have a vast aggregated repertoire, which is not yet the case for books. This confirms the initial analysis: the artistic fields face different challenges in terms of digitisation and require specific complementary approaches over and above the possible transversal rules.

The Open Science model was also discussed. It was made clear, however, that unlike Creative Commons, which offers an alternative contractual solution for intellectual property rights, the Open Science model makes research itself as well as research data freely accessible to everyone. This option excludes the granting of licences to determine the terms of use of a work. One possible solution, with regard to the use of data, would be to require platforms to operate in a similar way to Open Science in order to promote transparency.

A joint discussion between WG4 and WG1 then took place on the role of machines in intermediation. The salient points of this discussion concerned cultural diversity and the tools for making recommendations (and how these can be put into practice in public representations), the fair remuneration of authors and artists in a period leading to a paradigm shift with the emergence of AI, the different realities according to the cultural sectors concerned (which means that the generalist discourse is not appropriate), the need to improve existing systems (in particular collective societies) and to make known to politicians the major issues and different mechanisms that can support them.

To provide the best possible response to the above-mentioned issues, the two groups suggested the creation of a collective database (Creative Commons) on which the various music streaming platforms could make their algorithms accessible, in order to offer users greater transparency and alternatives in their consumption choices.

Recommendations

The following 24 proposals were put forward by participants:

1. The systematic payment of remuneration to authors and artists by Internet platforms and services negotiated by collective societies.

2. Encourage studies and experiments to broaden the remuneration base.

3. Remunerate/compensate AI output (in a similar way to private copying) or provide some sort of compensation for people who lose their jobs because of AI (more of a tax measure than a copyright measure).

4. License and receive remuneration on AI input, i.e. on reproductions of works used for input, in particular by means of extended collective licences (which will mean ensuring compatibility with Directive 2019/790).

5. A tax on revenues and an obligation on platforms and AIs to invest in national content.

6. Establishment of a European investment fund for artistic creation, along the lines of existing funds such as Eurimages ²² .

7. With regard to the investment model, support more open platforms in the spirit of the commons (e.g. Open Access or Creative Commons public platforms).

8. To the attention of governments, guarantee a full and fair sharing of value, integrating all layers (works, metadata, usage data and user data).

9. Initiate impact studies on the effects of AI and plan corrective and regulatory measures.

10. The creation of a generalised self-regulation system to experiment with new relationships between artists, collective societies, operators and the public.

11. Legally establish a common international register of usage data and personal data, an international body that would allow access to this data.

12. Strengthen international standard-setting bodies to improve remuneration for creators and give greater legitimacy to the use of collective management.

13. Recall that copyright is a human right (with constitutional status ²³ ) and recognise its professional nature.

14. This constitutional right should be linked to a right to transparency and access to data through data-sharing models.

15. Streaming platforms must be made accountable and transparent, and works must be revalued.

16. Credit (effective paternity and maternity rights) for rights holders.

17. Improving the networking of collective societies.

18. Tax the turnover of AI providers to compensate for the art effects generated by AI.

19. Legally guarantee (open) access to platform metadata for alternative recommendations.

20. Encourage research into disruptive ideas for alternative collective or individual remuneration models.

21. Support the creation of algorithms of general interest that promote cultural diversity and creative freedom.

22. A resale right on the increase in the stock market value of companies that use works to develop their activities.

23. Draw up a model extended collective agreement for negotiations with foreign countries.

24. Better coordination between employment law and copyright, as recommended for example in Bruno Racine's report ²⁴ .

To sum up, these proposals highlight the need to establish fairness in the sharing of the overall value generated by the use of works and services, to increase transparency in the platform data taken into account in collective management, to put in place innovative forms of governance to create common standards with binding effect, and to put in place international tax mechanisms to compensate for the imbalances that have arisen in the market (abuse of position by groups, economic parasitism, social dumping, etc.).

The reasons for the recommendations (rationals) are as follows: to promote cultural diversity and quality content, to improve the socio-professional situation of "artists" (a word that is still evolving), and to recognise, legally legitimise and promote high-quality (transparent, controlled) collective management.