Joint letter on AI to European Commission's Executive Vice-President Virkkunen and Commissioner Micallef

On 4 December 2024, 13 international and European organisations representing hundreds of thousands of writers, translators, journalists, performers, composers, songwriters, screen directors, screenwriters, visual artists, and other artists and creative workers, sent a letter to the European Commission's recently appointed Executive Vice-President Henna Virkkunen and Commissioner Glenn Micallef, who will be in charge of crucial EU policies related to generative Artificial Intelligence and copyright in the cultural and creative sectors. Find the letter below.

Brussels, 4 December 2024

Dear Executive Vice-President Virkkunen,

Dear Commissioner Micallef,

We are writing to you on behalf of a coalition of organisations representing the collective voice of hundreds of thousands of writers, translators, journalists, performers, composers, songwriters, screen directors, screenwriters, visual artists, and other artists and creative workers.

First, we would like to congratulate you on your appointment as Executive Vice-President of the European Commission for Tech Sovereignty, Security and Democracy and Commissioner for Intergenerational Fairness, Youth, Culture and Sport. As highlighted in your mission letters, your work will focus on “developing an AI strategy for cultural and creative industries” and considering “the need to further improve the copyright framework to address new challenges raised by market and technology developments”, while “improving the working conditions of artists and cultural professionals”. We welcome such policy objectives, and we look forward to collaborating with you to achieve these key priorities for the creative communities we represent.

For our members, AI represents an extraordinary technological advancement with immense potential to enhance various aspects of our lives, including in the cultural and creative sectors. However, as AI is deployed more and more widely, it is also important to acknowledge a darker aspect of this technology: all generative AI models in existence today have been trained in full opacity on enormous amounts of copyright-protected content and personal data, scraped and copied from the internet without any authorisation nor any remuneration for the creators we represent. In addition, the use of AI-generated deep fakes and other AI-manipulated content poses a significant threat to our democracies, our members’ reputation and moral rights, and to citizens’ trust in the veracity of digital content.

So far, the relevant EU legal framework (in particular the 2019 Directive on Copyright in the Digital Single Market, the General Data Protection Regulation, and the AI Act) governing the relationship between artificial intelligence, authors’ and performers’ rights, and data protection is too often misinterpreted, insufficiently applied, poorly enforced or simply ignored by generative AI models. In essence, it fails to adequately protect the rights of our creative communities and the value of their cultural works.

In sharp contrast to current practices, we firmly believe that authors and performers must have the right to decide whether their works can be used by generative AI and, if they consent, to be fairly remunerated. As you undertake the task of shaping the EU policy agenda for the next five years, we urge you to support a clearer and more efficient legal framework preserving the rights of creators and the integrity of their works. As representatives of the creative community in Europe, here is where we stand on this existential issue:

1.     Facing today’s reality and the flaws of the current EU Copyright framework in relation to generative AI

-        In 2019, the EU introduced an exception to copyright for text and data mining (TDM), provided that creators and other rightsholders have not expressly reserved their rights (Article 4 of the CDSM Directive) – without any impact assessment on the implementation nor effect of such provision.      

-        Although Article 4 was adopted years before the sudden rise of generative AI technologies and does not mention or define “Artificial Intelligence” and “Generative AI”, the European Commission appears to have retrospectively interpreted this exception broadly and without democratic overview to cover the systematic and extensive use of creators’ protected works and performances without any authorisation, despite the well-established principle of narrow interpretation of all exceptions in the EU acquis[1]. 

-        In the absence of relevant Court decisions, and as AI-generated outputs enter the market competing with human creations on unfair terms, it is hard to conceive how this broad interpretation would meet any public policy objectives and satisfy the three-step test, a fundamental safeguard enshrined in EU and international law intended to strike a fair balance between rightsholders and content users.        

-        In addition, more than five years after the adoption of Article 4 of the CDSM Directive, none of our members has been able to reserve their rights efficiently, and there is still significant uncertainty regarding the opt-out and how authors and performers can exercise it. As a result, generative AI providers have put the cart before the horse by using this exception without even giving a chance to creators to provide their consent and exercise their right of reservation. As of today, it would be quite naïve to believe that the tech and AI industries have not already scraped virtually all content available online and continue to do so on a daily basis – before any rules effectively come into force.        

-        Last but not least, as generative AI providers keep exploiting copyrighted works without transparency, it is virtually impossible for creators to know whether their works have been used and to take legal action against them, while generative AI systems available online and the widespread use of deepfakes are already having a major impact on our creative industries.

Such a situation is not only unacceptable for the authors and performers we represent but also prevents the development of an ethical and responsible AI ecosystem based on fair licensing terms. Therefore, we urge you to face today’s reality and critically assess the current legal framework applicable to copyright and generative AI. We firmly believe that consent, transparency and fair remuneration for authors and performers should be at the heart of the forthcoming AI strategy for the cultural and creative industries.

2.     If properly enforced, the AI Act will be a step in the right direction but will not be sufficient for a meaningful AI strategy for the cultural and creative industries

Our organisations welcomed the AI Act, in particular the requirement for providers of general-purpose AI models to comply with EU copyright law and publish sufficiently detailed information about the data used. We also strongly support the strengthening of transparency obligations around deep fakes and stress the importance of developing technical tools that can reliably and accurately differentiate authentic content from AI-generated, or manipulated, content.

Now that the Act must be concretely enforced, we urge you to ensure that the AI Office places transparency at the heart of its Code of Practice and template summary for AI providers. Without ambitious and detailed transparency obligations, it will be impossible for our members to know if their works have been used, to reserve their rights, and to avail themselves of the protection provided by EU law related to intellectual property, the protection of personal data or other relevant provisions. Since the tech and AI industries have already used our members’ works without any authorisation, it is no surprise that they are so adamant about the protection of trade secrets. In our view, all policy makers must clearly distinguish between legitimate trade secrets – where the input has been legally acquired – and the secretive use of copyrighted works to train AI models without any authorisation. A recipe may remain secret but that does not justify stealing the ingredients.

However, even with proper implementation, the AI Act will only serve as a temporary fix for a much larger problem unless an ambitious AI strategy for the cultural and creative industries is developed to ensure informed consent and remuneration for authors and performers.

3.     Key principles for the future AI strategy for the cultural and creative industries

In a context where the current EU legal framework is unenforced and unbalanced, does not protect our creative community, and ignores the specificities of the cultural and creative sectors, we call on you to directly address the profound disruption caused by the uptake of generative AI in these sectors, as well as its impact on creation and cultural diversity.

All future initiatives related to the use of AI for the cultural and creative industries must be based on a comprehensive and democratic debate leading to a clear legal framework preserving the rights and the integrity of creators’ works and personal data. It should address the many unresolved issues related to the TDM exception and clarify the terms of its extension to generative AI. In this debate, it is essential to consider that using works and performances in the context of AI models is radically different from other forms of exploitation. As the integrity of their work and their personal reputation may well be jeopardised by generative AI, authors and performers should retain the ability to consent or refuse such use of their work.

As AI-generated content draws much of its value from human creations exploited on a large scale, it is essential to consider effective and enforceable mechanisms to remunerate the creative community for the AI-generated output. Such mechanisms, however, must not operate to normalise or unduly encourage the replacement of human work by generative AI.

In conclusion, we urge you to place the principles of transparency, consent and remuneration of authors and performers at the heart of all future initiatives for an ethical and fair use of AI. We also look forward to collaborating with you and your teams to develop a framework advancing AI technologies in ways that serve and enhance human creativity, whilst continuing to promote original content and protecting the hundreds of thousands of authors and performers we represent.

[1] An important study commissioned by the German Authors‘ Rights Initiative states that the TDM-exception doesn't apply to GenAi: https://urheber.info/diskurs/executive-summary-english (summary in English) and https://urheber.info/diskurs/copyright-law-training-of-generative-ai (article in English).“