Europe Hits Pause on Its Own AI Rulebook

By the AI Boom Team | Published 20 November 2025 The EU AI Act delay is official: Brussels wants more time before its toughest AI rules bite. On 19 November, the European Commission published a “Digital Omnibus” package to simplify Europe’s tech laws. Its headline move pushes back the AI Act’s high-risk rules from August 2026 to December 2027.

What the EU AI Act delay covers

“High-risk” means AI used where mistakes can seriously hurt people. The postponed rules cover AI used in:

• job applications and exams;
• credit checks and health services;
• biometric identification and law enforcement;
• road traffic and utilities such as water and power.

The package also simplifies those annoying cookie pop-ups. More controversially, it would loosen GDPR so that tech companies can use Europeans’ personal data to train AI models, as Reuters reported.

Why Brussels blinked

Big Tech and the US government have pushed hard against Europe’s rules. Many businesses also complained the technical standards were not ready. The Commission insists that simplification is not deregulation. Critics are less sure, as Sifted explains.

What UK businesses should do

If you sell into the EU, or use AI to screen staff or customers, the rules are delayed, not dead. So use the breathing space. List where you use AI, note which uses could count as high-risk, and keep records. Our free AI audit worksheet makes that a one-afternoon job. For the UK angle, see why AI labs now want binding rules. Update, 2026: EU lawmakers have since agreed the delay. High-risk rules now apply to stand-alone AI systems from 2 December 2027 and to AI built into products from 2 August 2028, according to law firm Winston Taylor. Want to understand how AI is changing business, in plain English? Read Get AI Before It Gets You by the AI Boom team, or get in touch at hello@aiboom.uk. © aiboom.uk