Prohibited AI practices
Summary
AI practices that are prohibited in the Union, including subliminal manipulation, social scoring, and certain biometric identification uses. A new category (CSAM / non-consensual intimate imagery) was added by the Digital Omnibus on AI, effective 2 December 2026 — see amended_by field.
prohibitedunacceptable risksubliminalmanipulationsocial scoringbiometricfundamental rights
Read the full official text: https://ai-act-law.eu/article/5/
Track EU AI Act article 5 as evidence
eurocompliant maps this obligation to a checklist task and the evidence that satisfies it, alongside every other framework you follow.
Start free trial