Last updated:
A process to evaluate how AI systems affect fundamental rights (privacy, expression, non-discrimination) and identify mitigation measures.
A specialized impact assessment that maps AI use cases against human-rights frameworks (UDHR, ICCPR). It involves stakeholder consultations (including marginalized groups), scenario analysis of rights violations, and development of rights-preserving measures. Governance embeds HRIA findings into project requirements, contracts, and monitoring plans, and publicly publishes summary reports to ensure transparency and accountability.
Real world example:
A social-media company conducts an HRIA before deploying an avatar-generation AI. They consult disability advocates to avoid insensitive representations, implement bias checks for gender and race balance, and publish an HRIA summary showing how they mitigated potential harms to user dignity and expression.
