AI evaluating creditworthiness of natural persons. HIGH-RISK under Annex III 5(b).

📌 Annex III, point 5(b)

AI scoring individuals for consumer credit products. HIGH-RISK.

📌 Annex III, point 5(b)

AI scoring businesses (not natural persons). NOT high-risk - Annex III 5(b) covers only natural persons.

📌 Not covered

Multi-agent system (financial, ESG, sectoral, past decisions) producing a B2B/corporate risk opinion for credit officers. Human decides. NOT high-risk - Annex III 5(b) covers only natural persons.

📌 Not covered

AI-assisted loan application processing and approval for individuals. HIGH-RISK.

📌 Annex III, point 5(b)

AI making or influencing loan approval decisions for individuals. HIGH-RISK.

📌 Annex III, point 5(b)

AI determining interest rates. Context-dependent - may be high-risk if discriminatory impact.

📌 Art. 6(3)

AI for mortgage assessment and approval. HIGH-RISK under Annex III 5(b).

📌 Annex III, point 5(b)

AI determining credit limits for individuals. HIGH-RISK.

📌 Annex III, point 5(b)

AI assessing affordability for credit decisions. HIGH-RISK.

📌 Annex III, point 5(b)

AI prioritizing debt collection. May be high-risk if significantly impacts individuals.

📌 Art. 6(3)

AI for debt restructuring decisions. Context-dependent.

📌 Art. 6(3)

AI detecting fraudulent transactions. HIGH-RISK if can deny service access.

📌 Art. 6(3) / Annex III 5(b)

AI detecting credit/debit card fraud. Context-dependent.

📌 Art. 6(3)

AI detecting unauthorized account access.

📌 Art. 6(3)

AI detecting fraud in credit applications. If rejects applications → HIGH-RISK.

📌 Annex III, point 5(b)

AI for anti-money laundering. HIGH-RISK if denies account opening.

📌 Art. 6(3) / Annex III 5(b)

AI monitoring for suspicious transactions. Context-dependent.

📌 Art. 6(3)

AI scoring customer AML risk. May affect service access.

📌 Art. 6(3)

AI screening against sanctions lists.

📌 Art. 6(3)

AI screening for Politically Exposed Persons.

📌 Art. 6(3)

AI monitoring for suspicious activity patterns.

📌 Art. 6(3)

AI detecting market abuse and manipulation.

📌 Art. 6(3)

AI detecting MAR violations.

📌 Art. 6(3)

AI detecting potential insider trading.

📌 Art. 6(3)

AI automating regulatory report generation.

📌 Not listed

AI monitoring compliance with internal policies.

📌 Not listed

AI for automated trading. Not high-risk under AI Act (regulated under MiFID II).

📌 Not listed

AI for HFT. Regulated under MiFID II Art. 17.

📌 Not listed

AI providing investment advice. May be high-risk for retail if affects access.

📌 Art. 6(3)

AI investment advice for retail. Heightened suitability requirements.

📌 Art. 6(3)

AI investment advice for professionals. Lower protection requirements.

📌 Not listed

AI optimizing investment portfolios.

📌 Not listed

AI for automatic portfolio rebalancing.

📌 Not listed

AI achieving best execution for orders.

📌 Not listed

AI for liquidity provision.

📌 Not listed

AI routing orders to optimal venues.

📌 Not listed

AI for ESG analysis and scoring.

📌 Not listed

⚠️ HIGH-RISK: AI for life insurance pricing. Explicitly covered by Annex III 5(c).

📌 Annex III, point 5(c)

⚠️ HIGH-RISK: AI for health insurance pricing. Explicitly covered by Annex III 5(c).

📌 Annex III, point 5(c)

AI for property insurance pricing. NOT explicitly high-risk - context-dependent.

📌 Not listed

AI for motor insurance pricing. NOT explicitly high-risk - context-dependent.

📌 Not listed

AI for liability insurance pricing. NOT high-risk.

📌 Not listed

⚠️ HIGH-RISK: AI for life insurance underwriting.

📌 Annex III, point 5(c)

⚠️ HIGH-RISK: AI for health insurance underwriting.

📌 Annex III, point 5(c)

AI for property insurance underwriting. Context-dependent.

📌 Not listed

AI for claims assessment. May be high-risk if denies claims.

📌 Art. 6(3)

AI routing claims to appropriate handlers.

📌 Not listed

AI detecting fraudulent claims.

📌 Art. 6(3)

AI pricing based on driving behavior data.

📌 Art. 6(3)

⚠️ HIGH-RISK: AI filtering job applications. Annex III point 4(a).

📌 Annex III, point 4(a)

AI extracting data from CVs. May qualify for Art. 6(3) exemption if preparatory only.

📌 Art. 6(3)(d)

⚠️ HIGH-RISK: AI scoring/ranking candidates. Annex III point 4(a).

📌 Annex III, point 4(a)

⚠️ HIGH-RISK: AI matching candidates to roles.

📌 Annex III, point 4(a)

⚠️ HIGH-RISK: AI analyzing interviews.

📌 Annex III, point 4(a)

⚠️ HIGH-RISK: AI analyzing video interviews (may include emotion recognition).

📌 Annex III, point 4(a)

⚠️ HIGH-RISK: AI evaluating employee performance. Annex III point 4(b).

📌 Annex III, point 4(b)

⚠️ HIGH-RISK: AI predicting employee performance.

📌 Annex III, point 4(b)

⚠️ HIGH-RISK: AI influencing promotions. Annex III point 4(b).

📌 Annex III, point 4(b)

⚠️ HIGH-RISK: AI involved in termination decisions.

📌 Annex III, point 4(b)

⚠️ HIGH-RISK: AI monitoring employee behavior. Annex III point 4(c).

📌 Annex III, point 4(c)

⚠️ HIGH-RISK: AI monitoring productivity.

📌 Annex III, point 4(c)

⚠️ HIGH-RISK: AI allocating work tasks. Annex III point 4(c).

📌 Annex III, point 4(c)

AI forecasting workforce needs. Context-dependent.

📌 Art. 6(3)

AI analyzing compensation equity.

📌 Art. 6(3)

AI predicting employee flight risk.

📌 Art. 6(3)

AI recommending training courses.

📌 Not listed

AI chatbots for customers. LIMITED RISK - requires transparency disclosure.

📌 Art. 50(1)

AI chatbot providing advice. Art. 50 transparency required.

📌 Art. 50(1)

AI chatbot executing transactions. Art. 50 + service-specific requirements.

📌 Art. 50(1)

AI voice assistants. LIMITED RISK - requires transparency.

📌 Art. 50(1)

⚠️ Biometric identification - may be HIGH-RISK under Annex III point 1.

📌 Annex III, point 1

AI for customer verification. HIGH-RISK if can deny accounts.

📌 Annex III, point 5(b)

AI verifying customer identity. May involve biometrics.

📌 Art. 6(3)

AI segmenting customers for marketing.

📌 Not listed

AI predicting customer churn.

📌 Not listed

AI for product recommendations.

📌 Not listed

AI determining optimal customer interaction.

📌 Not listed

AI analyzing customer sentiment.

📌 Not listed

AI routing complaints to handlers.

📌 Not listed

AI extracting data from documents.

📌 Not listed

AI categorizing documents.

📌 Not listed

AI extracting text from images.

📌 Not listed

AI screening incoming emails.

📌 Not listed

AI analyzing contract terms.

📌 Not listed

AI reviewing contracts for issues.

📌 Not listed

AI-enhanced RPA.

📌 Not listed

AI for complex process automation.

📌 Not listed

AI monitoring data quality.

📌 Not listed

AI for automated reconciliation.

📌 Not listed

AI-enhanced internal capital models.

📌 Art. 6(3)

AI for Internal Ratings Based models.

📌 Art. 6(3)

AI for market risk assessment.

📌 Not listed

AI for Value at Risk calculations.

📌 Not listed

AI for portfolio credit risk. Context-dependent.

📌 Art. 6(3)

AI for Probability of Default.

📌 Art. 6(3)

AI for Loss Given Default.

📌 Not listed

AI for operational risk assessment.

📌 Not listed

AI for liquidity risk.

📌 Not listed

AI for climate risk modeling.

📌 Not listed

AI for stress test scenarios.

📌 Not listed

AI assisting model validation.

📌 Not listed

AI for access control decisions.

📌 Art. 6(3)

⚠️ Biometric ID may be HIGH-RISK under Annex III point 1.

📌 Annex III, point 1

⚠️ HIGH-RISK: Biometric identification. Annex III point 1.

📌 Annex III, point 1

AI analyzing behavioral patterns for auth.

📌 Art. 6(3)

AI detecting security anomalies.

📌 Not listed

AI detecting cyber threats.

📌 Not listed

AI for dynamic price adjustment.

📌 Art. 6(3)

AI for asset valuation.

📌 Not listed

AI valuing collateral. May affect credit decisions.

📌 Art. 6(3)

AI valuing properties. May affect mortgage decisions.

📌 Art. 6(3)

AI reviewing contracts, agreements, and legal documents. Analysis tool with lawyer oversight.

📌 Not listed

AI searching case law, precedents, and legal databases. Research tool for lawyers.

📌 Not listed

AI for electronic discovery in litigation. Document classification and relevance scoring.

📌 Not listed

AI-assisted contract generation using templates and clause libraries. Lawyer review required.

📌 Not listed

AI analyzing documents for M&A and transactions. Risk identification tool.

📌 Not listed

AI drafting legal briefs and memoranda. Lawyer review and approval required.

📌 Not listed

AI predicting litigation outcomes. HIGH-RISK if used by courts/affects access to justice. MINIMAL if internal law firm strategy tool.

📌 Art. 6(3) / Annex III 8

AI triaging legal inquiries and automating client intake workflow.

📌 Not listed

AI tracking billable hours and legal billing. Efficiency tool.

📌 Not listed

AI monitoring client compliance with regulations. Alert system.

📌 Not listed

AI preparing court filings. Administrative automation with lawyer oversight.

📌 Not listed

⚠️ HIGH-RISK: AI analyzing witness testimony. May affect legal outcomes under Annex III point 8.

📌 Annex III, point 8