Photo Credit: Getty Images
Meta has refused to sign the European Union's new artificial intelligence code of practice, calling the policy an "overreach" that threatens innovation and business growth.
"Europe is heading down the wrong path on AI," Meta's global affairs chief Joel Kaplan wrote Friday in a blunt LinkedIn post. The company, he added, had "carefully reviewed" the framework and concluded it imposes "legal uncertainties for model developers" while going "far beyond the scope of the AI Act."
The voluntary code, unveiled by the European Commission earlier this month, aims to prepare companies for full enforcement of the AI Act. That legislation, passed last year, introduces sweeping rules for transparency, safety, and documentation across all tiers of AI systems, particularly general-purpose models like those developed by Meta, OpenAI, and Google.
Kaplan, who took over Meta's policy portfolio earlier this year, said the company supports balanced regulation but believes the code's current form would "throttle the development and deployment of frontier AI models in Europe." Meta is not alone in its pushback. Tech giants including ASML and Airbus recently signed a letter urging Brussels to delay implementation by two years. OpenAI, however, has committed to signing the code despite earlier concerns about feasibility and legal clarity.
The Commission has so far refused to budge. The AI Act is set to begin phasing in from August 2, 2025, with a full compliance deadline for existing general-purpose AI providers set for August 2027. It bans certain "unacceptable risk" applications outright, including social scoring and cognitive manipulation, while imposing strict oversight on high-risk use cases like biometric surveillance and algorithmic hiring.
Under the code, AI developers must maintain up-to-date documentation, avoid training on pirated content, and honor opt-out requests from content creators. Critics argue the guidelines amount to a form of "soft enforcement" with unclear legal standing, one that could deter AI investment across the continent.
"We share concerns raised by others that this overreach will stunt European companies looking to build businesses on top of these models," Kaplan said.
The EU's digital chief, Thierry Breton, has emphasized that the guidelines are "a preparatory tool" and not a law. But as enforcement nears, tensions between regulators and the tech industry continue to escalate.
