Committee Procedure under AI Act
Follow topic LLM context A cited markdown file you can paste into your AI assistant (ChatGPT, Claude, a RAG or project knowledge base) to ground it in this topic. Contains: the overview, key law text, case law, enforcement and guidance for this topic. Everything links back to its source on overview.legal — legal information, not advice.The content specifically addresses 'Committee procedure' as a distinct procedural mechanism under the AI Act. This topic is not adequately covered by existing topics and requires its own dedicated entry to capture the specific procedural rules, voting mechanisms, composition requirements, and decision-making processes of regulatory committees established under the AI Act framework.
Overview
7 sources · Sep 25, 2026Legal Framework
The committee procedure under the AI Act governs how the European Commission adopts delegated and implementing acts with member state participation. The central procedural anchor is Article 97, referenced explicitly by Article 7 when the Commission amends Annex III to expand high-risk classifications:
"The Commission is empowered to adopt delegated acts in accordance with Article 97 to amend Annex III by adding or modifying use-cases of high-risk AI systems where both of the following conditions are fulfilled:"
— AI Act Art. 7(1)
This procedure binds the Commission to consult a committee of member state representatives before adopting measures that expand regulatory scope. The same procedural architecture extends into sectoral legislation amended by the AI Act. Article 106 inserts a new paragraph into Directive (EU) 2016/797, requiring that when the Commission adopts both delegated and implementing acts concerning AI safety components, the Chapter III, Section 2 requirements of the AI Act must be taken into account. Article 108 mirrors this across multiple provisions of Regulation (EU) 2018/1139, embedding AI Act compliance considerations into the committee-driven adoption of aviation safety rules.
The interplay is structural: Article 6 defines which AI systems qualify as high-risk safety components, and the committee procedure under Article 97 then determines how the Commission can expand or modify that classification through delegated acts.
Key Developments
No court rulings or DPA enforcement decisions have yet addressed the committee procedure under the AI Act specifically. The Act entered into force in August 2024, and the first delegated acts requiring committee consultation are anticipated as implementation progresses. The practical threshold for when the committee procedure is triggered turns on whether the Commission acts under a delegation clause—Article 7 for Annex III amendments, or the amended sectoral provisions in Articles 106 and 108.
A critical operational question concerns the substantive standard committees must apply when reviewing proposed acts. Article 7(3) sets evaluative criteria—including intended purpose, data processed, autonomy, and demonstrated harm—that committee members can use to assess whether a proposed delegated act is justified.
Status of the Debate
This topic is contested in doctrinal terms, though no court split is on record. The contestation arises from the intersection of the AI Act's committee procedure with pre-existing sectoral regulatory frameworks. Articles 106 and 108 amend established sectoral legislation, embedding AI Act requirements into committee deliberations that were originally designed for product safety assessments alone. Whether committee members in sectoral contexts (railway, aviation, machinery) possess sufficient AI expertise to evaluate Chapter III, Section 2 requirements meaningfully remains an open question.
The boundary dispute centers on whether the committee procedure functions as a substantive check—requiring genuine assessment of AI-specific requirements—or as a procedural formality where sectoral committees rubber-stamp Commission proposals. Resolution will likely come through the Court of Justice when a delegated act adopted under Article 97 is challenged for insufficient consideration of the Article 7(3) criteria, or when a sectoral implementing act is contested for failing to account for AI Act requirements under the amended provisions.
Practical Guidance
Map every delegated act trigger: Identify which provisions in your regulatory domain reference Article 97 or the amended sectoral articles, and track Commission announcements for upcoming delegated acts that could reclassify your AI systems as high-risk under Article 7.
Prepare Article 7(3) submissions: When the Commission consults on Annex III amendments, frame stakeholder input around the enumerated criteria—intended purpose, autonomy, demonstrated harm, and data sensitivity—to align with the committee's evaluative framework.
Monitor sectoral committee proceedings: For AI systems that are safety components under Article 6, track implementing act adoption under the amended provisions of Directive (EU) 2016/797 and Regulation (EU) 2018/1139, as these will incorporate AI Act Chapter III requirements.
Document conformity assessment readiness: Ensure your AI systems meet Chapter III, Section 2 requirements before sectoral committees adopt implementing acts that will apply those requirements, since the amended sectoral legislation makes compliance a precondition for market access.
Track committee composition and expertise: Engage with national representatives on relevant sectoral committees to ensure AI-specific concerns are raised during delegated and implementing act deliberations, particularly where sectoral expertise may not extend to AI risk assessment.
Nothing of this type on this topic.