Quick Reference
Risk Classification Framework
The AI Act classifies AI systems into four risk tiers. Each tier carries different obligations.
Banned outright. Includes social scoring, real-time biometric surveillance, and subliminal manipulation. Illegal from 2 February 2025.
Permitted with mandatory risk management, documentation, human oversight, and conformity assessment. Most obligations from August 2026.
Chatbots must disclose they are AI. Deepfakes must be labelled. AI-generated content must be disclosed.
No specific obligations under the Act. The vast majority of AI applications fall here. Voluntary codes of conduct encouraged.
Enforcement Timeline
The AI Act uses a phased enforcement schedule. Different obligations apply at different dates.
Act enters into force
Prohibited practices banned
GPAI & governance rules apply
High-risk AI rules apply
Annex I product rules apply
Full enforcement timeline with article references →