← Home

Annex III Use-Case Lookup

Search all high-risk AI use cases listed in Annex III and prohibited practices from Article 5. Find out if your AI system falls into a regulated category.

29 use cases found

Biometrics

Remote biometric identification systems

High-Risk

Facial recognition in public spaces, gait recognition, fingerprint matching for identification

Biometrics

Real-time remote biometric identification in public spaces (law enforcement)

Prohibited

Live CCTV facial recognition systems used by police or border guards in real-time

Biometrics

Biometric categorisation by sensitive attributes

Prohibited

AI inferring race, political opinions, religion, sexual orientation from biometric data

Biometrics

Emotion recognition systems

High-Risk

Detecting fatigue in drivers, stress monitoring in workplaces or classrooms

Critical Infrastructure

Safety components of critical digital infrastructure

High-Risk

AI managing electricity grids, water supply networks, internet exchange points

Critical Infrastructure

Road traffic management safety components

High-Risk

AI controlling traffic lights, autonomous driving safety systems, incident detection

Education & Vocational Training

AI determining access or admission to education

High-Risk

University admissions screening AI, school placement algorithms

Education & Vocational Training

AI assessing students or evaluating learning outcomes

High-Risk

Automated essay grading, exam monitoring AI, skills assessment tools

Employment & HR

AI for recruitment and selection of persons

High-Risk

CV screening AI, video interview analysis, automated job matching with profiling

Employment & HR

AI for employment conditions, promotion, or termination

High-Risk

Algorithmic performance management, AI deciding on pay raises or dismissals

Employment & HR

AI for task allocation and worker monitoring

High-Risk

Warehouse worker monitoring AI, gig economy dispatch algorithms, productivity scoring

Essential Services

Credit scoring and creditworthiness assessment

High-Risk

AI evaluating loan applications, mortgage eligibility, credit card limits

Essential Services

AI in healthcare and access to public benefits

High-Risk

AI triaging patients, AI deciding eligibility for social security, disability benefit assessment

Essential Services

Life and health insurance risk assessment

High-Risk

AI setting insurance premiums based on health or lifestyle data

Essential Services

Emergency call dispatch prioritisation

High-Risk

AI classifying 112/999 calls by urgency and routing emergency responders

Law Enforcement

Polygraphs and similar reliability-testing tools

High-Risk

AI-powered lie detection systems, deception analysis tools used in interrogations

Law Enforcement

Criminal risk assessment / recidivism prediction

High-Risk

AI tools predicting likelihood of reoffending to inform bail or sentencing decisions

Law Enforcement

AI profiling to predict criminal behaviour

Prohibited

Predictive policing systems based solely on profiling of persons

Law Enforcement

Evidence reliability assessment

High-Risk

AI assessing the reliability of witness statements or digital evidence in criminal proceedings

Migration, Asylum & Borders

Risk assessment of persons at borders

High-Risk

AI profiling travellers for risk at airports or land borders

Migration, Asylum & Borders

Asylum application assessment

High-Risk

AI supporting decisions on asylum claims, refugee status determination

Migration, Asylum & Borders

Travel document verification

High-Risk

AI-based passport and visa document authenticity checks

Justice & Democratic Processes

AI assisting courts in fact or law determination

High-Risk

AI tools helping judges analyse evidence, determine applicable law, or predict outcomes

Justice & Democratic Processes

AI influencing election outcomes or voter behaviour

High-Risk

Micro-targeting political ads, AI generating political campaign content at scale

Prohibited Practices (Art. 5)

Subliminal or manipulative techniques

Prohibited

AI embedding hidden cues in content to subliminally influence decisions without awareness

Prohibited Practices (Art. 5)

Exploitation of vulnerabilities (age, disability)

Prohibited

AI targeting children or disabled persons with manipulative persuasion techniques

Prohibited Practices (Art. 5)

Social scoring by public authorities

Prohibited

Government AI ranking citizens by their social behaviour and applying consequences

Prohibited Practices (Art. 5)

Facial recognition database scraping

Prohibited

AI building biometric databases from untargeted scraping of internet photos or CCTV

Prohibited Practices (Art. 5)

Emotion recognition at work or school

Prohibited

AI monitoring employee or student emotions via cameras or physiological sensors