What happened
On Wednesday, 16 March 2026, MEPs from the Internal Market and Civil Liberties committees approved a joint position that postpones the activation of certain rules on high‑risk artificial intelligence systems. The agreement, adopted by 101 votes in favour, aims to give companies more time before strict standards such as biometric or critical‑infrastructure AI must be compliant.
Why it matters
The delay means that EU businesses could still rely on sectoral safety rules for products already covered under the AI Act until 2 August 2028, buying them space to adapt without immediate penalties. Yet the postponement also signals a wobble in the Commission’s original timeline, raising concerns about regulatory certainty.
Context
The proposal builds on the EU AI Act that was meant to give consumers and regulators clear rules for high‑risk systems – such as those used in law enforcement, education or essential services. But as of now, key standards like watermarking AI media have not been finalised.
What comes next
The MEPs propose fixing the activation dates: 2 December 2027 for high‑risk AI systems listed in the regulation, and 2 August 2028 for safety‑component rules. They also want a ban on “nudifier” apps that create explicit images of real people without consent.
“We want predictable, stop‑the‑clock, simplified rules,” said Arba Kokalari (EPP, SE), co‑reporter for the Internal Market and Consumer Protection committee. “Simplifying the AI legislation will help companies scale up while protecting fundamental rights.”