Initiative overview
The EU AI Office was established by European Commission Decision on 24 January 2024, ahead of the formal entry into force of the EU AI Act (August 2024), in recognition of the need for a dedicated EU-level body to coordinate AI Act implementation and provide specialised technical and governance expertise. Its mandate is broader than enforcement alone: the AI Office serves as the EU's primary institutional voice on AI policy, representing the Commission in international AI forums including the OECD, G7, G20, the Council of Europe, ISO/IEC, and the International Network of AI Safety Institutes. The AI Office has four main functions. First, AI Act implementation and enforcement: it is the sole competent authority for supervising GPAI model providers across the EU, including conducting conformity assessments, requesting model evaluations, issuing warnings and, where necessary, imposing fines (up to EUR 15 million or 3% of global annual turnover for GPAI violations). Second, standards and guidance: it coordinates with European standards bodies (CEN/CENELEC, ETSI) on harmonised standards for AI Act compliance and publishes guidelines, codes of practice and technical documentation requirements. Third, the Office leads the multi-stakeholder drafting process for the GPAI Code of Practice, a key implementation instrument for GPAI model providers under Article 56 of the AI Act. Fourth, AI research and safety: the Office conducts and commissions research on AI safety risks, including systemic risks from GPAI models, and maintains the EU's scientific capacity for AI evaluation. It also manages the Apply AI Alliance under the Apply AI Strategy and coordinates AI Factories network development under InvestAI.
