EU Council and Parliament reach provisional deal to delay AI Act high-risk rules

Negotiators for the European Parliament and the Council of the EU reached a provisional political agreement on the Digital Omnibus on AI, a package of targeted amendments to the AI Act the European Commission had proposed the previous November after concluding that implementation of the high-risk rules was falling behind schedule. Rather than reopening the Act’s substance, the deal pushed back two compliance dates: obligations for standalone high-risk systems listed in Annex III — covering uses such as employment, education, biometrics and critical infrastructure — moved from 2 August 2026 to 2 December 2027, and obligations for high-risk AI embedded in products already covered by EU product-safety law, such as medical devices, moved from August 2027 to August 2028.

The package also introduced a new prohibition on AI-generated non-consensual intimate imagery and child sexual abuse material, a four-month grace period for watermarking requirements on systems already in use, and measures intended to reduce duplicated compliance work for AI embedded in regulated products.

The agreement was provisional: it still required formal endorsement by the full Parliament and final Council sign-off, which followed in the following two months, with formal adoption completed before the original August 2026 deadline the delay was designed to avoid triggering. The deal reflected a broader pattern in EU tech regulation through 2026 of loosening implementation timelines in response to industry pressure, without abandoning the underlying obligations.