The EU Just Delayed Its AI Act's Toughest Rules by 16 Months

The EU AI Act’s high-risk obligations were originally supposed to become binding on August 2, 2026. Right before that deadline, Europe passed an amendment called the “Digital Omnibus” that pushed the date back nearly a year and a half — the new deadline is December 2, 2027. But three other obligations took effect on the very same original date, and none of them got delayed.

What got delayed: the full compliance package for high-risk systems

The Digital Omnibus amendment formally entered into legal force on July 27, 2026, pushing the compliance deadline for standalone high-risk AI systems (the ones listed in Annex III) from August 2, 2026 to December 2, 2027. Take employment as an example: AI systems used for recruitment, candidate selection, performance evaluation, task allocation, monitoring of workers, and decisions on promotion or termination don’t have to immediately meet the full high-risk compliance package — risk management, data governance, technical documentation, human oversight, and the rest of it.

What didn’t get delayed: three transparency duties, right on schedule

That doesn’t mean nothing happened on August 2, 2026. Three transparency obligations under Article 50 of the AI Act — chatbot disclosure (interfaces must clearly tell users they’re interacting with AI), machine-readable marking of AI-generated content, and deepfake labeling — were not included in the delay at all, and took effect exactly as scheduled. The European Commission’s enforcement powers over general-purpose AI (GPAI) models also formally kicked in the same day.

Regulatory pace lagging technical pace is the norm here

Put the two pieces together and it reads like a pragmatic, staged approach: the compliance requirements that take companies real time and resources to meet — risk-management systems, technical documentation — got some breathing room, but the comparatively lightweight rule that most directly protects users, simply letting them know they’re talking to AI, didn’t budge by a single day. For companies building AI products, that means disclosure needs to be handled now, while the full high-risk compliance architecture still has more than a year of runway to prepare for.