Observed events
On 2 August 2026, the AI Act became generally applicable, with exceptions. The European Commission says the AI Office and national authorities began enforcing provisions then in application; certain transparency duties also started on that date.
The AI Omnibus entered into force on 27 July 2026. The Commission’s current timeline places Annex III high-risk use cases at 2 December 2027 and high-risk AI embedded in regulated products under Annex I at 2 August 2028.
Interpretation
The sequence points to a phased exposure map, not a blanket pause or one deadline for every system. A system’s intended purpose, the organisation’s role and the sector context all matter. This could mean relevant work in technology and data, employment, governance, intellectual property or product regulation—but public dates do not show that a company has a legal need, is non-compliant or has instructed counsel. This is Entriva’s interpretation of the official timetable.
Recommended actions
- Record each AI system, its intended purpose and the markets where it is used.
- Identify the organisation’s role in the AI value chain and whether a system falls within an Annex III use case or an Annex I product.
- Separate provisions already applicable from later high-risk dates; check other AI Act duties and sector rules as well.
- Confirm territorial scope and the national or EEA position against current official sources before advising on specific facts.
This EU-level overview is general information, not legal advice. It does not assess territorial scope or EEA incorporation. A possible legal need is only a hypothesis for a law firm to assess, not a confirmed mandate. The firm checks relevance and conflicts and decides whether and how to follow up.