Participation asymmetry in EU AI governance: a longitudinal analysis of consultation demographics and regulatory outcomes

Purpose The purpose of this study is to contribute to research on democratic governance in AI regulation by answering two questions: how did participation demographics shift across the EU AI Act’s consultation phases and what mechanisms explain the pattern; and how was asymmetric participation associated with the Act’s regulatory provisions. Design/methodology/approach This study uses an explanatory sequential mixed-methods design combining quantitative demographic analysis across three consultation phases (2020 White Paper, n = 1,215; 2021 legislative proposal, n = 303; and 2024 implementation guidance, n = 383) with qualitative provision-tracing through co-production theory and a four-indicator structural capture rubric. Findings EU citizen representation collapsed to 3.96% during the decisive legislative phase, while industry consolidated as the dominant stakeholder group across Phases 2 and 3. Three reinforcing mechanisms – resource disparities, institutional power imbalances and definitional flexibility – link demographic asymmetry to regulatory architecture. The Article 6(3) provider self-assessment framework satisfies all four structural capture indicators. The 2025 Digital Omnibus amendments, targeting the same accountability provisions, are consistent with the structural vulnerability embedded in this process. Research limitations/implications The analysis establishes institutional conditioning rather than deterministic causation, and findings from a single provision should not be generalised across the Act’s full architecture. Better Regulation consultation mechanisms appear structurally insufficient for technically complex domains and require reform toward deliberative mechanisms with binding democratic authority. Originality/value This study offers the first cross-phase longitudinal comparison of AI Act consultation demographics, introduces a replicable structural capture rubric for assessing participation quality and positions the AI Act as a diagnostic case for EU regulatory governance.