The European Parliament and the Council of the EU say a provisional Digital Omnibus agreement would push parts of the EU AI Act’s high-risk compliance timetable to December 2027 and August 2028, while rules already in force remain unaffected.
The European Parliament and the Council of the European Union have announced a provisional agreement to delay parts of the EU AI Act’s high-risk obligations under a Digital Omnibus simplification package.
According to the European Parliament, the provisional Digital Omnibus agreement would move the application date for high-risk AI obligations to 2 December 2027 for high-risk use cases and to 2 August 2028 for safety-component systems covered by EU sectoral legislation.
The Council of the European Union described the change in similar terms, saying high-risk AI provisions had been due to enter into force on 2 August 2026, but that the provisional agreement introduces later dates for two categories: 2 December 2027 for stand-alone systems and 2 August 2028 for embedded systems.
Naq Cyber, a compliance-focused publisher, also reported that high-risk obligations are being rescheduled under the Digital Omnibus, while duties already in force for prohibited AI practices and general-purpose AI governance are not affected by the high-risk delay.
The reported delay concerns the timetable for high-risk AI obligations. Naq Cyber says the rescheduling does not affect AI Act provisions that are already in force, including the ban on prohibited practices and governance requirements for general-purpose AI.
The European Parliament’s announcement also links the provisional deal to broader simplification measures and a ban on “nudifier” apps. The Parliament’s summary says the agreement is part of a package intended to simplify rules while adjusting the compliance calendar for high-risk AI systems.
The AI Act applies obligations in stages, so the compliance impact depends on the type of AI system and the role of the organisation deploying or supplying it. The Council’s statement distinguishes between stand-alone high-risk systems and embedded systems, with separate dates for each category under the provisional agreement.
For organisations planning AI Act compliance, the practical effect is that some high-risk obligations may have a longer implementation window than previously expected. However, the sources do not indicate that all AI Act duties are delayed. Businesses still need to assess whether they are subject to provisions that are already active or unaffected by the revised high-risk schedule.
Both the European Parliament and the Council describe the measure as a provisional agreement. That means organisations should treat the revised dates as an agreed political position reported by EU institutions, while continuing to monitor the formal legislative process and final text.
For now, the clearest source-backed timeline is that high-risk AI obligations previously expected in August 2026 would shift to 2 December 2027 for stand-alone or high-risk use-case systems and 2 August 2028 for embedded safety-component systems covered by sectoral EU law, according to the Council and Parliament announcements.
The European Parliament and the Council of the European Union have announced a provisional agreement to delay parts of the EU AI Act’s high risk obligations under a Digital Omnibus simplification package.
What does not appear to change The reported delay concerns the timetable for high risk AI obligations.
Naq Cyber says the rescheduling does not affect AI Act provisions that are already in force, including the ban on prohibited practices and governance requirements for general purpose AI.
Continue reading
