Meta declared it will not participate in the EU’s voluntary Code of Practice for general-purpose AI models, warning it introduces “legal uncertainties” and exceeds the boundaries of the EU AI Act.
Published on July 10, the code requires transparency on training data, adherence to copyright rules, and safety assessments. Meta’s Chief Global Affairs Officer Joel Kaplan asserted that Europe is “heading down the wrong path,” arguing compliance could “throttle the development and deployment of frontier AI models” within the region . While signing the code offers reduced administrative burden and clarity, non-signatories like Meta may face heightened regulatory scrutiny as the AI Act takes full effect on August 2, 2025
