Home / The Act / Chapter IX / Article 82

Article 82 — Compliant AI systems which present a risk

APPLIES AUG 2, 2026
Chapter IX · Governance
Source: EUR-Lex · CELEX 32024R1689
Verified 2026-07-16
PLAIN-LANGUAGE SUMMARY — ORIGINAL INTERPRETATION, NOT THE OFFICIAL TEXT

If an evaluation shows a high-risk system complies with the AI Act yet still endangers health, safety, fundamental rights or other public interests, the authority must require the operator to eliminate that risk within a set period. Corrective action must reach every affected system on the Union market, and the Commission and other Member States are informed at once. The Commission then reviews the measure, decides whether it is justified, and may propose further steps.

OFFICIAL TEXT Read verbatim on EUR-Lex ↗
The authoritative wording lives on EUR-Lex, quoted there in all 24 languages. We deep-link to the paragraph level rather than reproducing the full text. The consolidated post-Omnibus version will be linked the day its CELEX publishes.
AT A GLANCE
Chapter: IX — Post-market monitoring, information sharing and market surveillance
Applies: Aug 2, 2026
Tier: Governance
RELATED
How this is enforced →Risk tiers →Related obligations →
PENALTY EXPOSURE
—
of worldwide turnover — see Art. 99 penalties framework.
← PREVIOUS · ART. 81 Union safeguard procedure NEXT · ART. 83 → Formal non-compliance
ELSEWHERE IN CHAPTER IX — Post-market monitoring, information sharing and market surveillance
Art. 72 Post-market monitoring by providers and post-m Art. 73 Reporting of serious incidents Art. 74 Market surveillance and control of AI systems Art. 75 Mutual assistance, market surveillance and con Art. 76 Supervision of testing in real world condition Art. 77 Powers of authorities protecting fundamental r Art. 78 Confidentiality Art. 79 Procedure at national level for dealing with A