AI Board

EU Advisory body under Articles 65–67 of the AI Act, composed of one senior representative per Member State and supported by the AI Office as secretariat. Its mandate spans issuing harmonisation guidelines, resolving cross-border jurisdictional conflicts, coordinating joint enforcement actions, and advising the AI Office on GPAI systemic-risk classification and codes-of-practice adequacy.

Semantic Classification

Content

• Advisory body composed of representatives from EU Member States, supported by the AI Office, with the mission of coordinating uniform AI Act implementation, issuing guidelines, and facilitating cooperation among national authorities.

  Source

  Primary: EU AI Act Articles 65-67 Context: Member State coordination mechanism

  Regulatory Context

  The AI Board serves as the central coordination mechanism ensuring consistent AI Act application across all 27 EU Member States, preventing regulatory fragmentation and facilitating knowledge exchange among national authorities.

  Composition (Article 65)

  Member State Representatives

  Each EU Member State appoints:

• One member to AI Board

• One alternate for member absence

  Level: Senior representatives from national competent authorities with AI expertise

  European Data Protection Supervisor (EDPS)

  Observer status to ensure AI Act-GDPR coherence

  Secretariat

  AI Office provides secretariat functions:

• Administrative support

• Meeting organisation

• Documentation preparation

• Implementation of Board decisions

  Functions and Tasks (Article 66)

  1. Guidelines and Recommendations

  Practical Implementation Guidance

• Risk classification: Interpretation of Annex III use cases

• Conformity assessment: Best practices for procedures

• Technical documentation: Templates and standards

• Human oversight: Implementation methodologies

  Harmonisation Opinions

• Common specifications proposals (Article 41)

• Harmonised standards evaluation (Article 40)

• Systemic risk criteria interpretation (Annex XIII)

  Targeted Advice

• AI Office: On GPAI supervision approaches

• Commission: On legislative amendments

• Member States: On enforcement strategies

  2. Cross-Border Coordination

  Jurisdictional Conflicts

  Resolve disputes when:

• Multiple Member States claim jurisdiction

• Differing interpretations arise

• Coordinated enforcement needed

  Information Exchange

  Facilitate sharing of:

• Enforcement experiences: Successful and unsuccessful actions

• Technical findings: Model evaluations, testing results

• Emerging risks: New AI capabilities, unforeseen harms

  Joint Actions

  Coordinate:

• Multi-state investigations: GPAI providers operating EU-wide

• Harmonised enforcement: Consistent penalties, corrective measures

• Crisis response: Serious incident management

  3. Best Practice Development

  Market Surveillance Excellence

• Identify effective surveillance methodologies

• Develop inspection protocols

• Create authority training programmes

  Regulatory Sandbox Learnings

• Share sandbox outcomes across Member States

• Identify innovation-friendly approaches

• Recommend regulatory adaptations

  FRIA Templates (Article 27)

• Develop fundamental rights impact assessment methodologies

• Share public authority experiences

• Create sectoral guidance (law enforcement, education, healthcare)

  4. AI Office Support

  GPAI Oversight

  Advise AI Office on:

• Systemic risk model classification (Article 51)

• Enforcement priorities

• Codes of practice approval (Article 56)

  Strategic Planning

• Annual enforcement priorities

• Resource allocation recommendations

• International cooperation strategies

  Decision-Making (Article 66)

  Voting Rules

• Simple majority: Most decisions

• Qualified majority: Classification of systemic risk models (Article 51)

  Quorum

  Majority of Member State representatives must be present.

  Transparency

• Meeting agendas published in advance

• Decisions and opinions published (excluding confidential information)

  Working Methods

  Meetings

• Regular sessions: At least quarterly

• Extraordinary sessions: Upon AI Office or Member State request

• Virtual participation: Enabled for efficiency

  Working Groups

  AI Board may establish:

• Thematic groups: By sector (healthcare, finance, law enforcement)

• Technical groups: By topic (biometric systems, GPAI, conformity assessment)

• Ad hoc groups: For specific urgent issues

  Annual Work Programme

  AI Board adopts programme covering:

• Priority topics

• Planned guidelines

• Stakeholder consultations

• Cooperation initiatives

  Key Deliverables

  Opinions and Guidelines

  Published on:

• High-risk AI system classification edge cases

• Substantial modification interpretation (Article 28)

• Deployer obligations in specific sectors

• GPAI transparency information standards (Annex XII)

  Annual Report

  AI Board contributes to AI Office annual report on:

• Uniform implementation progress

• Persistent divergences among Member States

• Recommendations for legislative clarification

  Enforcement Coordination

• Common enforcement priorities

• Coordinated investigations outcomes

• Penalty consistency guidance

  Relationship to Other Bodies

  AI Office

• Secretariat support: AI Office provides administrative resources

• Implementation: AI Office executes AI Board recommendations

• Collaboration: Joint working groups, consultations

  Scientific Panel of Independent Experts (Article 68)

• Technical input: Panel advises Board on GPAI technical issues

• Alerts: Panel qualified alerts inform Board discussions

• Coordination: Joint consideration of systemic risks

  Advisory Forum (Article 67)

• Stakeholder perspectives: Advisory Forum feeds into Board deliberations

• Consultation: Board seeks Advisory Forum input on guidelines

• Transparency: Advisory Forum ensures Board considers diverse views

  National Competent Authorities

• Board members: National authorities represented on Board

• Implementation: Authorities implement Board guidelines domestically

• Reporting: Authorities report back to Board on outcomes

  Sectoral Engagement

  Financial Services

  Coordinate with:

• European Banking Authority (EBA)

• European Securities and Markets Authority (ESMA)

• European Insurance and Occupational Pensions Authority (EIOPA)

  Healthcare

  Coordinate with:

• European Medicines Agency (EMA)

• Medical Device Coordination Group (MDCG)

  Consumer Protection

  Coordinate with:

• Consumer Protection Cooperation Network

• Product Safety Network

  Data Protection

  Coordinate with:

• European Data Protection Board (EDPB)

• National data protection authorities

  Strategic Priorities (Expected)

  Phase 1 (2024-2026): Foundation

• High-risk AI classification guidance

• Conformity assessment harmonisation

• Market surveillance capacity building

  Phase 2 (2027-2029): Maturation

• GPAI supervision coordination

• Cross-border enforcement

• Regulatory sandbox integration

  Phase 3 (2030+): Evolution

• AI Act amendment recommendations

• International regulatory alignment

• Emerging technology adaptation

  Member State Benefits

  Consistent Application

• Avoid regulatory fragmentation

• Predictable compliance environment

• Level playing field for EU AI industry

  Shared Expertise

• Access to other Member States’ experiences

• Technical capacity building

• Reduced duplication of effort

  EU Voice Projection

• Coordinated international positions

• Enhanced global AI governance influence

• Regulatory export potential

  Transparency and Stakeholder Engagement

  Public Consultations

  AI Board conducts consultations on:

• Draft guidelines

• Classification opinions

• Enforcement approaches

  Stakeholder Dialogues

  Regular engagement with:

• Industry: Compliance challenges, innovation concerns

• Civil society: Fundamental rights, consumer protection

• Academia: Research findings, technical developments

• Standardisation bodies: Standards development coordination

  Publication

  AI Board publishes:

• Adopted opinions and guidelines

• Annual work programmes

• Meeting summaries (excluding confidential matters)

  Challenges and Solutions

  Challenge: Divergent National Interpretations

  Solution: Rapid opinion issuance on contentious issues, common Q&A repository

  Challenge: Varying Enforcement Resources

  Solution: Capacity building programmes, shared technical tools, joint investigations

  Challenge: Technological Pace

  Solution: Regular review cycles, Scientific Panel input, regulatory sandbox learnings

Provenance