EU pushes AI Act's high-risk compliance deadline from August 2026 to December 2027

The Digital Omnibus on AI, passed in June, resets the clock for Annex III high-risk systems — but obligations still apply retroactively to anything already deployed.

Alex Rivera

Editor

EU pushes AI Act's high-risk compliance deadline from August 2026 to December 2027

The EU AI Act's original enforcement date for high-risk system obligations — August 2, 2026, covering provider requirements under Articles 9–17 and deployer requirements under Article 26 — has been pushed back. Following a European Parliament vote on June 16, the Digital Omnibus on AI proposal became law on June 29, 2026, resetting the compliance clock for the high-risk tier.

Under the revised timeline, obligations for Annex III high-risk AI systems (the use-case-based category covering things like employment screening, credit scoring, and biometric identification) move from August 2, 2026 to December 2, 2027. Sectors that were already regulated before the AI Act — like healthcare, where high-risk AI is embedded in medical devices — get until August 2, 2028.

What doesn't change

The extension is breathing room, not an exemption. Obligations still apply to any high-risk system that was already in place before August 2, 2026 — deployers don't get to reset their own clock by having launched early. For teams building or deploying high-risk AI in EU markets, the practical move is to keep the compliance program on track for the new dates rather than treating the delay as the obligations going away.

Source: Holland & Knight. Travers Smith

X / TwitterLinkedIn

This space is available

Advertise your product to our AI-focused audience.

Advertise here

More from AI News