Skip to content

Answers

When do the EU AI Act's high-risk rules apply?

the short answer

After the Digital Omnibus changes, obligations for stand-alone high-risk AI systems (Annex III) apply from 2 December 2027, and for AI built into regulated products (Annex I) from 2 August 2028. The deadlines moved; the duties, including human oversight and record-keeping, did not.

George McKenna, Chairman and CTO, ETT Group. Updated 15 October 2026.

The original dates were 2 August 2026 for Annex III systems and 2 August 2027 for Annex I. A provisional agreement on the Digital Omnibus was reached in May 2026, and the changes have since entered into force.

For a business deploying AI in areas such as credit, insurance pricing, employment or access to essential services, the extra time is best used to build the evidence the Act will expect: what the system did, under whose oversight, and on what data. Retrofitting that later is harder and more expensive than designing it in now.

This is a summary, not legal advice. Check the current position with your legal team.

Sources: Gibson Dunn on the Omnibus agreement · Hunton on the Omnibus entering into force

Show Your Working

A fortnightly piece on building and buying AI that pays its way. What I'm building, what it costs, and the questions to ask before you sign anything.