EU AI Act — Prohibited Practices (Article 5): AI Compliance Requirements
EU AI Act Article 5 bans eight categories of AI outright — no grace period, no exceptions. These prohibitions became enforceable February 2, 2025 and are subject to the HIGHEST penalties in the entire AI Act. Key bans: (1) Social scoring of individuals by public/private bodies. (2) AI systems exploiting vulnerabilities (age, disability, socioeconomic status) to manipulate behavior. (3) Subliminal AI techniques that distort behavior. (4) Emotion recognition in workplaces and educational institutions. (5) Biometric categorization to deduce race, political opinion, religion, sexual orientation. (6) AI systems predicting crime risk based on personality or profiling. (7) Real-time remote biometric identification in publicly accessible spaces for law enforcement (narrow exceptions). (8) Scraping facial images from internet/CCTV to build recognition databases. If your AI product touches any of these — you have ZERO compliance buffer. Stop immediately.
Summary of publicly-available regulatory text as of 2026-08-22. Verify against current official sources before relying on this for compliance decisions. Not legal advice.
Key Facts
February 2, 2025
August 2, 2025
€35,000,000 or 7% of global annual turnover — highest penalty in the entire AI Act
What Your Business Must Do
4 compliance requirements identified. Critical requirements carry the highest risk of enforcement action.
Prohibited Practices Self-Audit
CriticalConduct an immediate audit of all AI systems against EU AI Act Article 5. Document your review. If any system falls under a prohibited category, cease use immediately. Even a low-risk appearance is not a defense — Article 5 is strict liability.
Deadline: August 2, 2025
AI Act Art. 5Emotion Recognition Workplace Check
CriticalIf you use any AI system that infers emotional state from video/audio/biometrics in your workplace or educational setting — this is NOW prohibited. Includes: performance monitoring AI, call center emotion scoring, student engagement AI. Remove or stop using such systems.
Deadline: August 2, 2025
AI Act Art. 5(1)(f)Social Scoring Prohibition Check
High PriorityAny AI system that evaluates trustworthiness, social behavior, or "social credit" to determine treatment in unrelated contexts is banned. Includes insurance AI using spending behavior for unrelated coverage decisions. Document that your AI does NOT perform social scoring.
Deadline: August 2, 2025
AI Act Art. 5(1)(c)AI Literacy Obligation (Article 4)
Medium PriorityAlso effective February 2, 2025: all providers and deployers of AI systems must ensure "sufficient AI literacy" of their staff. This is a soft but real obligation. Implement AI training for employees who interact with, manage, or oversee AI systems.
Deadline: February 2, 2025
AI Act Art. 4Who Does This Apply To?
Applies to: any provider, deployer, importer, or distributor that places on the EU market, puts into service, or uses an AI system in the EU that falls into one of the eight prohibited categories of Article 5 — regardless of company size or where the provider is established, if the system or its output is used in the Union. There is no grace period and no SME exemption: the prohibitions have been enforceable since 2 February 2025. The banned practices include social scoring, exploitation of vulnerabilities (age, disability, socioeconomic status), subliminal manipulation that distorts behaviour, emotion recognition in workplaces and educational institutions, biometric categorisation inferring protected characteristics, predictive-policing based solely on profiling, untargeted scraping of facial images to build recognition databases, and (with narrow law-enforcement exceptions) real-time remote biometric identification in publicly accessible spaces. If a product touches any prohibited category it must be withdrawn — there is no compliance pathway. Separately, the Article 4 AI-literacy obligation applies to all providers and deployers. Highest penalty tier in the Act: EUR 35,000,000 or 7% of global annual turnover.
Recent Enforcement Actions
Against:
Recent Regulatory Guidance
European Commission — Guidelines on prohibited AI practices under Article 5 of the AI Act (2025)
European Commission published binding guidelines (February 2025) clarifying Article 5's eight prohibited categories: (1) subliminal techniques distorting behavior; (2) exploitation of vulnerabilities (age, disability, socioeconomic); (3) social scoring by public/private bodies; (4) crime-risk profiling based on personality; (5) untargeted facial-image scraping for recognition databases; (6) emotion recognition in workplaces and education; (7) biometric categorization to deduce protected characteristics; (8) real-time remote biometric identification in public spaces (narrow LE exceptions). Each prohibition was operationalized with specific examples of in-scope AI systems and the documentary evidence supervisory authorities expect from operators.
Key Case Law & Precedent
CNIL v. Clearview AI (Oct 2022)
Commission nationale de l'informatique et des libertés (France) · 2022France CNIL fined Clearview AI €20 million for unlawful biometric processing — scraping facial images from the internet to build a facial-recognition database without lawful basis. The decision is the leading European precedent on Article 5(1)(e)'s prohibition on untargeted facial-image scraping for recognition databases. CNIL's reasoning is now embedded in the Commission's prohibited-practices guidelines and represents the de facto enforcement standard across all EU national supervisory authorities for the same prohibition.
Outcome: €20M fine, order to cease processing French residents' facial data, deletion order; €5.2M additional periodic-penalty fine in 2023
Case referenceIndustry Playbooks covering EU AI Act — Prohibited Practices (Article 5)
These industry playbooks include jurisdiction-specific checklist items and guidance for EU AI Act — Prohibited Practices (Article 5).
Frequently Asked Questions
Does EU AI Act — Prohibited Practices (Article 5) apply to my business?
EU AI Act Article 5 bans eight categories of AI outright — no grace period, no exceptions. These prohibitions became enforceable February 2, 2025 and are subject to the HIGHEST penalties in the entire AI Act. Key bans: (1) Social scoring of… Use Aegis Firma's free scanner to get a personalized assessment in under 5 minutes.
What is the penalty for non-compliance?
The maximum penalty under EU AI Act — Prohibited Practices (Article 5) is: €35,000,000 or 7% of global annual turnover — highest penalty in the entire AI Act. Fines are typically scaled by company size, severity of violation, and whether violations were willful or accidental.
How do I comply with EU AI Act — Prohibited Practices (Article 5)?
The 4 requirements above cover the core obligations. The fastest path to compliance is: (1) conduct an AI risk assessment, (2) document your AI systems, (3) implement transparency disclosures where required. Aegis Firma generates all required documents automatically.
Official Source
https://artificialintelligenceact.eu/article/5/Last updated: 2026-08-22 — verify at source before relying on this information.
Don't leave compliance to chance
Aegis Firma scans your AI tools, tells you exactly which regulations apply, and generates all required documents — in 30 minutes.
Start your free compliance scan