High-Risk Checker

Remote biometric ID, biometric categorisation, or emotion recognition (outside the contexts already prohibited)

e.g. Face-ID access control above a reliability threshold, gait-recognition systems, voice-print identification· Annex III §1

AI used as a safety component in critical infrastructure (digital, road, water, gas, power)

e.g. AI controlling grid frequency, traffic-light optimisation that affects safety· Annex III §2

AI used to determine access, evaluate learning, or assign people to education or vocational training

e.g. Admission scoring, automated essay grading at scale, AI-driven proctoring decisions· Annex III §3

AI used in recruitment, candidate scoring, performance evaluation, or worker management

e.g. CV-ranking software, AI interview scoring, gig-platform task allocation, AI-driven performance reviews· Annex III §4

AI used to determine access to essential public or private services

e.g. Credit scoring for individuals, insurance-risk pricing for natural persons, AI-driven welfare eligibility, emergency dispatch prioritisation· Annex III §5

AI used by law enforcement for risk assessment, evidence evaluation, or polygraph

e.g. Risk-of-reoffending scoring, AI-assisted truth detection· Annex III §6

AI used in migration, asylum, border control, or visa decisions

e.g. Risk-of-irregular-migration scoring, AI-assisted asylum interviews· Annex III §7

AI used to assist judicial decision-making or in elections/democratic processes

e.g. AI legal-research assistants used by judges to weigh evidence, voter-targeting AI in elections· Annex III §8

None of the above describe my system

e.g. Marketing chatbots, AI content generation, recommender systems, AI translation, general internal AI tooling