European Union: European Commission published third draft of General-Purpose Artificial Intelligence Code of Practice

On 11 March 2025, the European Commission published the third draft of the General-Purpose Artificial Intelligence (AI) Code of Practice. The third draft, developed through collaboration among industry, academia, and civil society, amends previous versions by streamlining structure, clarifying provisions, and enhancing commitments. The Code applies to providers of general-purpose AI models and those classified as having systemic risk. The obligations are structured across four working groups focusing on transparency, copyright, risk assessment, and risk mitigation. It aligns with the European Union's AI Act and international approaches, ensuring proportionality to risks, model provider size, and evolving AI governance needs. Notably, it strengthens reporting commitments, introduces a Model Documentation Form for transparency, and includes an Appendix with recommendations for the AI Office. It also outlines transparency and copyright obligations. Providers must maintain up-to-date model documentation to support downstream AI system developers and, upon request, provide information to the AI Office or national authorities for compliance assessments. The copyright section mandates a policy ensuring compliance with EU copyright law, addressing content mining, infringement risks, and complaint handling. The safety and security section also provides a framework for compliance with AI Act rules on General-Purpose AI models with systemic risk.