EU’s General-Purpose AI Code of Practice
The General-Purpose AI (GPAI) Code of Practice was published on 10 July 2025. It aims to help AI industry players comply with the European Union’s AI Act. This voluntary framework addresses safety, transparency, and copyright obligations for providers of general-purpose AI models. Member States and the European Commission are currently reviewing the Code’s adequacy. Following endorsement, AI providers who adopt the Code can demonstrate legal compliance more easily and reduce administrative burdens.
Purpose and Scope of the GPAI Code of Practice
The Code assists providers of general-purpose AI models to meet legal duties under the AI Act. It is designed through a multi-stakeholder process involving independent experts. The Code is voluntary but offers legal certainty and easier compliance proof. It focuses on three main areas – transparency, copyright, and safety and security. The latter applies specifically to providers of highly advanced AI models with systemic risks.
Transparency Chapter
This chapter provides a Model Documentation Form. It helps AI providers disclose essential information about their models. The form is user-friendly and supports compliance with Article 53 of the AI Act. Transparency ensures users and regulators understand how AI models function and their potential impacts. It promotes trust and accountability in AI deployment.
Copyright Chapter
Providers must comply with EU copyright laws when developing and deploying AI models. The Copyright chapter offers practical policies and solutions to meet these obligations. This includes respecting intellectual property rights and managing data sources used for training AI. The chapter guides providers in avoiding copyright infringements and legal disputes.
Safety and Security Chapter
This chapter targets providers of the most advanced general-purpose AI models that may pose systemic risks. It outlines state-of-the-art risk management practices. These include safety measures to prevent harm and security protocols to protect against misuse or attacks. Compliance with Article 55 of the AI Act is supported by this chapter. It ensures that high-risk AI systems operate responsibly.
Implementation and Impact
Once endorsed by the EU Member States and Commission, the Code will serve as a benchmark for AI providers. Signing the Code signals commitment to ethical and legal AI practices. It streamlines regulatory processes and encourages innovation by clarifying compliance pathways. The Code complements upcoming Commission guidelines on general-purpose AI concepts.
| Related | |
|---|---|
