EU GPAI Code
EU General-Purpose AI Code of Practice
Voluntary code enabling general-purpose AI model providers to demonstrate compliance with EU AI Act GPAI obligations. Three chapters cover transparency (model documentation), copyright compliance, and safety/security for systemic-risk models. Adherence creates legal presumption of conformity.
Jurisdiction
European Union
Enacted
Aug 1, 2025
Effective
Aug 2, 2025
Enforcement
EU AI Office and national competent authorities
Published July 10, 2025; formally approved August 1, 2025; GPAI obligations in force August 2, 2025
European Commission — Digital StrategyWhy It Matters
Establishes the practical compliance standard for general-purpose AI model providers under the EU AI Act, covering transparency, copyright, and safety obligations. Creates a legal presumption of conformity for signatories.
Recent Developments
Signatory Taskforce established to facilitate coherent application. Signatories include Anthropic, OpenAI, Google, Microsoft, Amazon. GPAI Template for data disclosure published July 24, 2025.
At a Glance
Applies to
Who Must Comply
- General-purpose AI model providers (estimated 5-15 globally)
- Providers of GPAI models with systemic risk (safety chapter)
Obligations fall on:
Applicability thresholds:
Safety Provisions
- Model Documentation Form with 10-year retention requirement
- Copyright compliance policies using state-of-the-art identification technologies
- Safety frameworks and incident reporting for systemic-risk models
- Information sharing with downstream providers and AI Office
Exemptions
Open Source Exemption
GPAI models released under free and open-source licenses are exempt from transparency obligations unless the model poses systemic risk
- • Released under free and open-source license
- • Does not pose systemic risk
Compliance & Enforcement
Key Dates
Aug 2, 2025
GPAI model provider obligations under EU AI Act take effect
Aug 2, 2026
End of enforcement grace period for GPAI obligations
Penalties
€15M or 3% revenue (whichever higher)
View on map
European Union
Focus Areas
Cite This
APA
European Union. (2025). EU General-Purpose AI Code of Practice.
Related Regulations
EU CRA
Mandatory cybersecurity requirements for all products with digital elements placed on the EU market, including AI software. Requires security by design, vulnerability handling, incident reporting to ENISA, software bills of materials, and CE marking for market access.
EU AI Content Transparency Code
Voluntary code of practice under the EU AI Act that sets out how providers of generative AI systems can meet the Article 50(2) duty to mark AI-generated or manipulated content in a machine-readable, detectable way, and how deployers can meet the Article 50(4) duty to label deepfakes and AI-generated text published on matters of public interest. The Commission and the AI Board have assessed it as an adequate means of demonstrating compliance.
Italy D.Lgs. 160/2026
First of two legislative decrees implementing Italy's AI law (Law 132/2025) and adapting national law to the EU AI Act. Title I sets the first organic rules for police use of AI systems on a human-oversight basis, including exceptional, time-limited and judicially authorised real-time biometric identification and a ban on biometric databases built by untargeted mass scraping of the web. Title II introduces Penal Code Article 437-bis, punishing failure to adopt security measures in high-risk AI systems and their unlawful alteration, and gives claimants procedural tools for damages caused by AI systems: access to the system's technical documentation, a presumption of causal link, an alternative forum near the claimant's residence and direct action against the insurer.
DE Digital Violence Bill (GgdG)
German government bill creating a standalone Digital Violence Act (GgdG) that gives victims court-ordered disclosure of anonymous offenders' identities, data preservation, content removal and temporary account suspension against platforms, and adding Criminal Code offences for sexualised deepfakes, reputation-damaging deepfakes, identity misuse and covert tracking.
Norway AI-in-Schools Recommendations
Norwegian government national recommendations directing that generative AI should, as a general rule, not be used by pupils in primary school (grades 1-7), with gradual, teacher-supervised use permitted in lower secondary (grades 8-10) and appropriate-use education in upper secondary. Issued through the Directorate for Education and Training (Udir).
FR SREN
France's 2024 "digital space" law strengthening national digital regulation and enforcement levers via ARCOM across platform safety and integrity issues.
Last updated March 9, 2026. Verify against primary sources before relying on this information.