Europe’s General-Purpose AI Code of Practice lands as a voluntary compliance pathway under the AI Act

[Founding Archive] AI20 editorial coverage based on European Commission (Shaping Europe’s digital future) (2025-07-10). On 10 July 2025, the European Commission announced that it had received the final General-Purpose AI Code of Practice — a voluntary instrument developed by 13 independent experts with input from over 1,000 stakeholders, including model providers, SMEs, academics, safety experts, rightsholders and civil society. The Code is designed to help industry comply with AI Act rules on general-purpose AI that would enter into application on 2 August 2025, with AI Office enforceability phased in later for new and existing models. Structurally, the Code has three chapters: Transparency and Copyright (for all GPAI providers) and Safety and Security (for providers of the most advanced models). Once endorsed by Member States and the Commission, voluntary signatories can demonstrate relevant AI Act compliance by adhering to the Code, with the Commission stating that this path offers reduced administrative burden and increased legal certainty compared with proving compliance by other means. Why it matters: Soft-law instruments often become the de facto operating standard before hard enforcement arrives. Early engagement with the Code’s transparency, copyright and safety chapters can de-risk market access and align multi-national assurance programmes. AI20 perspective: Multi-stakeholder codes are alliance-native tools. AI20 should track signatory practice, share implementation lessons across members, and keep civil-society and SME perspectives visible so “compliance pathways” do not become large-lab monocultures. (Category: Responsible AI; Region: European Union; Tags: Code of Practice; GPAI; AI Office; transparency; safety and security. Original source: https://digital-strategy.ec.europa.eu/en/news/general-purpose-ai-code-practice-now-available)