Commission Decision Establishing the European AI Office

• Authority: European Union
• Jurisdiction: Multinational
• Collections: Multinational
• Most recent activity: Enacted
• Most recent activity date: 2024-02-21

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• Governance failure

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• AI system security vulnerabilities
• Competitive dynamics
• AI pursuing its own goals
• AI possessing dangerous capabilities
• Lack of capability or robustness

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has minimal direct coverage of specific risk subdomains (2-3 subdomains), with implicit references to governance structures (6.5), system safety monitoring (7.1, 7.2, 7.3), and potential mentions of security vulnerabilities (2.2). The document primarily establishes administrative structures rather than addressing specific AI risks and harms. Most coverage is procedural rather than substantive risk mitigation.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Information
• Public Administration (excl. National Security)

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the sector are described.

• Scientific Research & Development Services
• Educational Services
• National Security

All other sectors had no coverage in this document.

Sectors Summary

This document does not govern specific economic sectors but rather establishes a cross-sectoral governance body (the European AI Office) that will oversee AI regulation across all sectors. The Office operates in the Information sector as part of the EU's digital governance infrastructure and will coordinate with sector-specific authorities across multiple industries.

Legislative Status

Hard Law

Legislative Status Summary

This is a binding Commission Decision establishing an administrative office with specific legal authority, enforcement powers, and mandatory tasks under EU law. The document uses mandatory language throughout and creates legally binding obligations.

Actors

Proposer

• Entity Types: AI_Governance_Actor
• Summary: The document is a Commission Decision issued by the European Commission, with Commissioner Thierry Breton signing on behalf of the Commission. The Commission has the legal authority under EU treaties to establish such administrative structures.

Target

• Entity Types: AI_Developer, AI_Deployer
• Summary: The document targets providers of AI models and systems who will be subject to monitoring, evaluation, and investigation by the AI Office. Specific focus on general-purpose AI model providers and those developing systems with systemic risks.

Enforcer

• Entity Types: AI_Governance_Actor
• Summary: The AI Office is explicitly tasked with enforcement activities including investigating infringements, monitoring implementation, and coordinating with national authorities. The Commission retains decision-making authority while the Office provides investigative and preparatory support.

Monitor

• Entity Types: AI_Governance_Actor
• Summary: The AI Office has extensive monitoring responsibilities including tracking AI market evolution, monitoring emergence of risks, monitoring implementation of rules, and coordinating with various advisory and oversight bodies. Multiple governance bodies are established for ongoing monitoring and evaluation.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document addresses multiple AI lifecycle stages with primary focus on Verify and Validate (evaluation tools and methodologies), Deploy (monitoring implementation), and Operate and Monitor (ongoing risk monitoring and enforcement). It also covers Plan and Design through regulatory sandboxes and guidance development, but does not substantially address data collection or model building stages.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• AI Systems
• General Purpose AI
• Open Weight / Open Source

Technical Scope Summary

The document explicitly mentions AI models, AI systems, and general-purpose AI models with particular focus on very large general-purpose AI models with systemic risks. It does not explicitly define compute thresholds, frontier AI, task-specific AI, foundation models, generative AI, predictive AI, or open-source models, though it does reference cooperation with the open-source community.