Proposing PaFRIA: A Participatory Fundamental Rights Impact Assessment Aligned with the EU AI Act

Under the EU AI Act, assessing the impact of AI systems on fundamental rights (FRIA) is a core obligation for certain high-risk AI system deployers (Article 27). To support compliance, the AI Office is tasked with developing a FRIA template. However, no such template currently exists. AI Act Recital 96 explains the FRIA obligations and suggests stakeholder involvement in FRIA. To address the gap in guidance and support stakeholder involvement as suggested in Recital 96, we propose PaFRIA, a participatory FRIA in line with the AI Act. It assists AI deployers in complying with the AI Act's FRIA obligations, and facilitates stakeholder involvement in the assessment process. We iteratively refined PaFRIA through participatory engagement with deployers, affected end-users, and experts. The iterations revealed six clusters of tensions, which we translated into basic principles for designing a participatory AI Act-aligned FRIA. We present the resulting PaFRIA and discuss its implementation.