Meta has announced it will not sign the European Union’s new Code of Practice for general-purpose AI (GPAI) models, calling the framework legally ambiguous and claiming it goes beyond the scope of the AI Act. The decision sets Meta apart from other major AI developers such as OpenAI, Anthropic, and France’s Mistral AI, who have stated they will support the voluntary Code.
In a LinkedIn post published on Friday, 18 JUly, Meta’s Chief Global Affairs Officer Joel Kaplan wrote, “Europe is heading down the wrong path on AI… 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.”
Meta’s stance adds fuel to a broader debate over the EU’s regulatory approach to AI. Earlier this month, over 40 European companies, including ASML and Airbus, signed an open letter urging the Commission to “Stop the Clock” on implementing the AI Act, warning that overregulation could stunt innovation and undermine Europe’s ability to compete with China and the United States.
Commission: Code is voluntary, but a benchmark
The European Commission told EU Perspectives that while the Code of Practice is not mandatory, it provides an important baseline for responsible AI development.
If a provider decides not to sign the Code of Practice, it will have to demonstrate other means of compliance. – Thomas Reigner, Commission spokesperson
“All GPAI providers will have to comply with the AI Act as of 2 August 2025,” said Commission spokesperson Thomas Reigner. “The Code of Practice is a voluntary tool, but a solid benchmark. If a provider decides not to sign the Code of Practice, it will have to demonstrate other means of compliance,” added spokesperson.
The Commission also warned that companies opting out may face closer scrutiny: “Companies who choose to comply via other means may be exposed to more regulatory scrutiny by the AI Office.”
Meta is, so far, the only major AI company to publicly decline the Code. While others, such as OpenAI and Mistral, have embraced it as a positive step toward transparency and accountability, Meta argues that the document introduces “overreach” and will “throttle the development and deployment of frontier AI models in Europe.”
You might be interested
What comes next
The final version of the Code was published by the European Commission on July 10, after a long drafting process. Companies with GPAI models already on the market must sign before August 2 if they wish to benefit from the “presumed compliance” status the Code provides under the AI Act.
While Meta’s decision not to sign may appear symbolic, it also raises real questions about enforcement and industry alignment. With the AI Act’s key obligations for GPAI developers taking effect in August 2025, the Commission has signalled it will be watching closely how companies choose to demonstrate compliance.
As Europe moves to set a global standard on AI regulation, the coming months will test whether its voluntary instruments can win buy-in from the very companies they aim to guide, or whether regulatory fragmentation will begin to take hold.
