TL;DR
- Meta has declined to sign the EU’s AI Code of Practice just weeks before enforcement.
- The company cites legal uncertainties and overreach in the voluntary code.
- Meta warns that the rules will stunt AI innovation and harm European developers.
- The AI Act takes effect August 2, 2025, with full compliance required by 2027.
- Other major AI firms like OpenAI, Google, Microsoft, and Mistral AI are also pushing back.
Meta has refused to sign the European Union’s Code of Practice for general-purpose AI (GPAI) models, just weeks ahead of the EU AI Act entering into force.
“Europe is heading down the wrong path on AI,” wrote Joel Kaplan, Meta’s Chief Global Affairs Officer, in a LinkedIn post. “We have carefully reviewed the European Commission’s Code of Practice for GPAI models, and Meta won’t be signing it.”
Kaplan argues that the code introduces legal uncertainties for AI model developers and imposes obligations beyond the AI Act’s intended scope.
What the EU Code Requires
The EU Code of Practice — released earlier this month as a voluntary compliance framework — asks companies to:
- Regularly document and update their AI systems.
- Ban training on pirated or unauthorized content.
- Honor requests from copyright holders to exclude their work from datasets.
Kaplan calls these terms regulatory overreach and warns that they will “throttle the development and deployment of frontier AI models in Europe.”
About the AI Act
The EU AI Act is the world’s first comprehensive legislation governing AI. It uses a risk-based framework to classify AI applications as:
- Unacceptable risk: e.g., cognitive manipulation or social scoring (outright banned).
- High risk: includes uses in biometrics, facial recognition, education, and employment.
For high-risk applications, developers must register systems, ensure transparency, and maintain risk management standards.
Industry Pushback and EU Response
Major tech firms — including Alphabet, OpenAI, Microsoft, and Mistral AI — have urged the Commission to delay enforcement. However, the European Commission has stood firm.
On Friday, the EU also published final compliance guidelines for developers ahead of the August 2, 2025 implementation.
What’s Next?
Companies with general-purpose AI models already in the market — such as Meta’s Llama, Google’s Gemini, or Anthropic’s Claude — must achieve full compliance by August 2, 2027.
As enforcement looms, it remains to be seen whether Meta’s refusal will influence other firms or lead to regulatory escalation.
The Data
| Regulation Detail | Summary |
| AI Act Effective Date | August 2, 2025 |
| Compliance Deadline | August 2, 2027 |
| Scope | Applies to all GPAI systems operating in EU |
| Key Companies Affected | Meta, OpenAI, Google, Microsoft, Mistral AI |
