Skip to main content
Regulation4 min read

AI Omnibus: Rebaseline Controls Without Stopping

The communicated timelines for high-risk systems change planning, but do not remove the need for documented preparation.

Updated
Conceptual editorial illustration for “AI Omnibus: Rebaseline Controls Without Pausing Preparation”.
Conceptual image generated for this analysis; it does not document a real event.AI-generated conceptual image by Velqory.

In brief

The Commission page on the AI Omnibus says the package entered into force on 27 July 2026 and communicates application dates for two high-risk system categories [claim-reg-omnibus]. The prudent decision is to rebaseline milestones, budget, and evidence without treating an extended timeline as an exemption from preparation [claim-reg-planeamento].

  • Replan using the communicated dates for the stated categories.
  • Keep preparation and control evidence in place.
  • Review the plan when legislation or classification changes.

Communicated timeline

The Commission page states 2 December 2027 for Annex III high-risk systems and 2 August 2028 for high-risk AI embedded in physical products under Annex I [claim-reg-omnibus].

Management decision

It is an editorial recommendation to use the communicated dates to rebaseline the programme while retaining inventory, ownership, and validation work [claim-reg-planeamento].

Scope and review

This analysis is limited to the European Union and the Commission’s published description. Concrete classification and legal effect require case-specific assessment; review if the legislative text or use case changes [claim-reg-omnibus].

Consulted on the stated update date.

  1. 01AI Omnibus enters into force | Shaping Europe’s digital futureShaping Europe’s digital future · EU-DIGITAL
  2. 02Commission starts enforcing AI Act rules and new transparency requirements on 2 August | Shaping Europe’s digital futureShaping Europe’s digital future · EU-DIGITAL