On 10 July, the European Commission received the final version of the General-Purpose AI Code of Practice, a voluntary framework developed by 13 independent experts with input from over 1,000 stakeholders, including AI providers, SMEs, academics, civil society, and AI safety experts.
The Code is intended to support industry in preparing for the AI Act’s rules on general-purpose AI, which take effect on 2 August 2025. Enforcement will begin one year later for new models and two years later for existing models.
Structured around three chapters—transparency, copyright, and safety and security—the Code offers practical tools and guidance. It includes a model documentation form to simplify transparency obligations, policy recommendations for compliance with EU copyright law, and best practices for managing systemic risks associated with the most advanced models.
CEPIS welcomes this development as a significant step toward building a trustworthy and transparent AI ecosystem in Europe. CEPIS encourages all applicable companies and organisations within its network to sign the Code. Doing so will not only support a healthy and responsible AI landscape but also bring tangible benefits in the form of legal certainty and reduced compliance burdens.
Once endorsed by EU Member States and the Commission, providers that voluntarily sign the Code will be able to demonstrate compliance with relevant provisions of the AI Act. The Code will be complemented by forthcoming Commission guidelines that clarify the scope of general-purpose AI obligations under the AI Act.
