EU Releases General-Purpose AI Code Of Practice

On July 10, 2025, the European Commission published the General-Purpose AI Code of Practice (the Code), which offers voluntary guidance for general-purpose artificial intelligence (AI) model providers to meet certain obligations under the European Union Artificial Intelligence Act (AI Act). The Code is intended to reduce administrative burden on these AI providers and offer more legal certainty than if they were to prove compliance through alternative methods.

The Code was developed through a multi-stakeholder process involving independent experts and stakeholders from diverse backgrounds, including AI model providers, small and medium-sized enterprises, industry organisations, academia, and civil society. The Code is composed of three chapters:

1. Transparency. Aimed at allowing AI providers to document required information to comply with their transparency obligations under Article 53 and the corresponding Annexes XI and XII of the AI Act, including providing a Model Documentation Form that acts as a user-friendly tool for AI providers to compile this information.
2. Copyright. Provides practical guidance for regulated entities to address point (c) of Article 53(1) of the AI Act, which requires AI providers to have a policy in place that complies with the European Union’s copyright law and related rights.
3. Safety and Security. Outlines practices for AI providers of advanced AI models, which must manage systemic risks in accordance with Article 55 of the AI Act.

In the weeks following publication, the Code will be evaluated by European Union Member States and the European Commission to determine its adequacy. Following endorsement by these parties, AI providers can sign the Code and agree to comply with its obligations. The European Commission also plans to release guidelines on key concepts for general-purpose AI models to complement the Code.

Summary By: Imtiaz Karamat