Meta Rejects EU’s AI Code of Practice; OpenAI and Anthropic Open to Sign
- Martina
- 25 July 2025, Friday
In a not-so-surprising move, the company Meta Platforms, which owns Facebook and Instagram, has refused to sign the EU’s voluntary AI Code of Practice. Meanwhile, OpenAI and Anthropic are willing to follow the international guidelines.
Meta Says the EU’s Code of Practice and AI Act Restrict AI Developers
Published on July 10, 2025, the EU’s General-Purpose AI Code of Practice is designed to guide the industry through the rules and regulations of the upcoming EU AI Act—a comprehensive legal framework for artificial intelligence within the European Union that was passed last year and will take effect on August 2, 2025.
The Code of Practice—ultimately a voluntary document—was created by 10 independent experts with the input from over 1,000 relevant stakeholders, including AI safety specialists, academics, model providers, small- and medium-sized enterprises, and civil society organizations.
Across three chapters—Transparency, Copyrights, and Safety and Security—the voluntary AI framework sets rules and limits on how AI developers can collect copyrighted content and requires companies to develop policies for addressing copyright laws. The Code also mandates that model developers assess and reduce systematic risks associated with general-purpose AI models, such as the development of chemical and biological weapons or issues related to losing control over the model.
Currently, providers of general-purpose AI models are encouraged to sign the Code of Practice and demonstrate compliance with the relevant AI Act regulations by following the Code. However, Meta Platforms, an important player in the industry, has refused to sign the AI obligations.
“We have carefully reviewed the European Commission’s Code of Practice for general-purpose AI (GPAI) models and Meta won’t be signing it. This Code introduces a number of legal uncertainties for model developers, as well as measures which go far beyond the scope of the AI Act,” wrote Joel Kaplan, Meta’s Chief Global Affairs Offer, in a statement.
As the EU has repeatedly emphasized in its press release, the Code of Practice is not legally enforceable and is described as a “voluntary tool” intended for guidance and regulation in advance of the AI Act. It is, however, also stated that those AI providers who sign the Code will benefit from a lower administrative burden and greater legal certainty compared to those who do not. Companies that fail to sign the document will also face higher regulatory scrutiny.
Unlike Meta, other leading AI model providers such as OpenAI, Anthropic, or Microsoft have stated their intention to sign the Code of Practice. This is despite some of them having previously been accused of illegally reproducing copyrighted content or using pirate sites to download large volumes of copyrighted material.
For example, Anthropic is currently being sued by music publishers over claims that its Claude chatbot reproduced copyrighted lyrics. Previously, the company also faced a lawsuit over the use of copyrighted books to train AI. The court ultimately ruled the latter case as “fair use.”
“We believe the Code advances the principles of transparency, safety and accountability – values that have long been championed by Anthropic for frontier AI development. If thoughtfully implemented, the EU AI Act and Code will enable Europe to harness the most significant technology of our time to power innovation and competitiveness,” Anthropic stated on its website.
Meta and EU Companies Stand Against the AI Act
Meta’s reluctance to sign the EU Code doesn’t come as a total surprise. In 2023, the company faced a record $1.3 billion fine over Facebook data transfers to the US. This year, it was also fined €200 million for violating the EU’s Digital Markets Act (DMA) with Facebook and Instagram’s “pay or consent” model.
As far as AI is concerned, the multinational tech company has been skeptical of the EU’s approach to AI and the regulations in the AI Act — instead, it has openly supported the Trump administration’s lenient stance on tech regulations.
Meta’s Kaplan also recently took to LinkedIn to shine a spotlight on the letter sent to the European Commission, in which more than 40 EU company leaders opposed the AI Act.
Companies across various industries, including Airbus, E-ON, Mercedes-Benz, Siemens, and Spotify, have stated that the legislative framework puts AI development at risk of being “disrupted by unclear, overlapping and increasingly complex EU regulations.” According to them, the AI Act is likely to put developers at a disadvantage and delay the development and deployment of AI technologies across businesses, thus making Europe less competitive.
“We share concerns raised by these businesses that this over-reach will throttle the development and deployment of frontier AI models in Europe, and stunt European companies looking to build businesses on top of them,” Kaplan seconded.
What Does the EU AI Act Mean for Musicians and the Rest of the Music Business?
An important part of the AI Act is an ‘opt-out’ system for AI development, which means that copyright holders who do not want their content to be used in training AI models and technologies must explicitly specify so.
As a result, major record labels like Sony Music Group and Warner Music Group have sent letters to prominent AI firms to inform them that they do not consent to having their music or lyrics used in AI training.
The debate over AI companies using artists' work worldwide to train their models without permission has grown significantly over the years. Of course, most of these developers firmly believe that using copyrighted content in this way should be considered “fair use.” It’s important to note, however, that using copyrighted art and music involves more than just lyrics and melodies (though those would be enough); it also includes artists’ voices and likenesses.
Within the EU, AI laws prevent AI companies from training their models on copyrighted materials if rightsholders have simply “opted out.” Under the EU AI Act, providers of general-purpose AI models must not only follow these laws and regulations but also demonstrate their compliance by showing transparency during their training processes.
This rule should apply even in cases where AI models have been trained outside the EU. Naturally, this could have a profound impact on AI companies worldwide, especially those that train their models in countries where AI laws are less stringent or non-existent, such as the US.
Many artists and music companies have been vocal about the importance of implementing proper AI laws that protect musicians and their art. While the passing of the EU AI Act in 2024 is a significant step, it was reported that many industry insiders are concerned that the transparency obligations in the ACT are being “watered down in implementation.“
As a result, a broad coalition of organizations in Europe’s cultural and creative sectors, representing creators and rightsholders across music, literature, film, and more, has been established to actively advocate for artists’ rights and protection.
Through the ‘Stay True To The Act’ website, the coalition has hosted videos from, among others, more than 30 European music creators, urging EU policymakers to properly implement the copyright and transparency requirements in the AI Act. These creators include Portuguese artists Tomás Wallenstein and Dela Marmy, Poland’s Eurovision contender Justyna Steczkowska, Mihail from Romania, or Finnish musicians Ellinoora and ABREU.
“Passed in early 2024, the AI Act has provided authors, performers, and other rightsholders with tools that are essential to exercise and enforce their rights,“ states the website. “Our coalition is concerned that the implementation standard is being brought to a level so low that it would undermine the objectives of the AI Act, contravene EU law, and ignore the intention of the EU legislator.“
With the AI Act coming into effect in August (actually, in just a few days), it’s only a matter of time before we see how effective the AI regulations are in practice and implementation.
The European Union has been emphasizing the need for clear AI laws for a long time. So, let’s hope that the AI Act will provide real support to artists and creators and effectively challenge companies that refuse to follow these rules and regulations—like Meta, for instance.
Ready to get your music out there?
Distribute your music to the widest range of streaming platforms and shops worldwide.
