Publishers, Journalists, and Film Producers Challenge Europe’s New AI Law
A broad coalition, including journalists, publishers, and notable figures from the film and music industries, has raised significant concerns regarding the latest draft of proposed regulations aimed at ensuring compliance with Europe’s new AI Act. This coalition, representing major copyright holders, argues that the current draft fails to adequately protect their rights under both the AI Act and European Union copyright law.
Key Concerns of the Coalition
The statement from this influential group adds pressure on the European Union to reconsider or reject the “code of practice” intended to guide large AI companies in complying with the EU AI Act. The timing of this intervention is critical, as it follows warnings from key lawmakers about the dilution of requirements for AI companies to assess their models for “systemic risks” to society and fundamental rights.
Former negotiators expressed that the current wording of the draft contradicts the intent of the AI Act itself, potentially allowing for serious harm to elections and the overall European economy.
Implications for Copyright Holders
The new statement emphasizes that the draft code “sets the bar so low” that it provides insufficient support for authors, performers, and other rights holders in enforcing their rights when AI companies’ web crawlers encounter their copyrighted content. The AI Act mandates that AI companies must respect the wishes of rights holders concerning the use of their works for training AI models.
This legislation, which is the most comprehensive of its kind globally, is set to come into effect for providers of “general purpose AI” (GPAI) models, such as OpenAI’s GPT-4o, in early August. As the deadline approaches, the industry urgently anticipates clearer guidance. The code of practice, currently in its third draft, is being developed by external experts for the European Commission’s new AI Office, with the final draft expected in May.
Criticism of the Draft Code
The rights-holder groups assert that “no code would be better than the fundamentally flawed third draft.” This coalition includes prominent organizations such as the International Federation of Film Producers’ Associations, the Federation of European Publishers, the International Federation of the Phonographic Industry, and the International Federation of Journalists.
One critical issue for rights holders is that the draft code does not provide options for rejecting the use of their copyrighted works for AI training. The only mechanism recognized is the often-ignored robots.txt protocol, which allows website operators to request that web scrapers refrain from accessing certain content. Moreover, the draft fails to offer AI companies any concrete guidance on how to comply with such requests.
Transparency and Compliance Issues
The latest draft exacerbates the situation by removing requirements for AI companies to be transparent about their compliance with rights holders’ requests to protect their content from scraping. “A key objective of the AI Act is to equip authors, performers, and other rights holders with tools to exercise and enforce their rights,” the coalition stated, stressing that the third draft represents a regression from this goal.
In response, the European Commission reiterated its stance, indicating that the code is a compromise and that it will assess the final draft before making a decision on its approval.
The American Copyright Debate
In the U.S., the current administration shows little indication of proposing comprehensive regulations akin to the EU AI Act. However, there is increasing pressure from the industry to establish federal rules to preempt conflicting state-level AI legislation and provide more legal certainty, particularly concerning copyright issues.
Recent evidence in ongoing lawsuits, such as those involving authors Sarah Silverman and Ta-Nehisi Coates against Meta, has revealed that Meta used a library of pirated books for AI training. The contentious EU draft states that AI companies need only make “reasonable efforts” to avoid crawling piracy websites.
The current administration is developing an “AI action plan” and has solicited industry feedback. Organizations like OpenAI have suggested that the U.S. should provide AI companies with clear copyright exemptions, arguing that failing to do so would disadvantage American firms against their Chinese counterparts, who do not adhere to copyright laws.
In contrast, Thomson Reuters, which recently won a case against a company that improperly replicated its content for an AI service, opposes OpenAI’s proposals. Their chief product officer emphasized the importance of respecting copyright in AI development, highlighting the significant intellectual investment that goes into creating content.
The British government is also contemplating a copyright exemption for AI companies, provoking backlash from the creative sectors. Technology Secretary Peter Kyle urged that musicians and filmmakers should embrace change rather than resist it.
