The European Commission has today received the final version of the Code of Conduct for General-Purpose Artificial Intelligence, prepared by 13 independent experts in collaboration with more than 1,000 stakeholders from industry, academia, civil society, as well as experts in security and copyright.
The Code serves as a voluntary tool to support the industry in aligning with the provisions of the Artificial Intelligence Act (AI Act), which comes into force on August 2, 2025. The application of the rules will become mandatory one year later for new models, and within two years for existing ones. The goal is to ensure the safe and transparent use of advanced AI models on the European market.
The Code is divided into three chapters:
- Transparency – includes a template for easy documentation of models,
- Copyright – provides guidelines for compliance with EU legislation in this field,
- Safety and Protection – outlines best practices for managing systemic risks, intended only for developers of the most advanced models.
Signatories of the Code will be able to more easily demonstrate compliance with the requirements of the AI Act, benefiting from reduced administrative burden and greater legal certainty. The Code will be supplemented by additional guidance and recommendations from the Commission, which will be published before the legal obligations take effect.
