Overview of the AI Act Omnibus Developments
The European Commission’s Digital Omnibus on AI proposals has experienced a turbulent negotiation process, culminating in the breakdown of the 28 April trilogue. This article examines the key issues, stakeholder positions, and potential implications for AI governance professionals.
Key Negotiation Points
High‑risk AI System Deadlines
Two critical deadlines have been set: 2 Dec 2027 for standalone Annex III high‑risk systems and 2 Aug 2028 for AI embedded in regulated products. The imminent enforcement deadline of 2 Aug 2026 for high‑risk systems adds pressure to reach a timely agreement.
Intersection with Existing EU Digital Rules
Debates focus on whether AI systems that fall under existing sectoral safety legislation—such as medical devices, industrial machinery, toys, and connected cars—should be exempt from the AI Act or remain governed solely by those sectoral rules. The European Parliament advocates for carve‑outs, while the Council and Commission resist, aiming to preserve the AI Act’s horizontal framework.
Annex I Product Coverage
The list of products covered by harmonised EU safety legislation (Annex I) is a contentious issue. Removing certain high‑risk AI applications from the AI Act’s scope could shift regulatory responsibility to sector‑specific regulations, potentially leading to a “deregulatory” outcome rather than simplification, as warned by MEP Michael McNamara and civil‑society groups.
Stakeholder Perspectives
Parliamentary Position
The Parliament seeks to reduce overlapping obligations by carving out AI applications from the AI Act, especially in sensitive sectors like financial supervision, law enforcement, and border management.
Council and Commission Stance
Both bodies oppose broad carve‑outs to maintain a unified regulatory architecture, fearing fragmentation and loss of the AI Act’s overarching principles.
Standardisation Community
Sebastian Hallensleben of CEN‑CENELEC emphasises the importance of preserving the standards framework that underpins high‑risk AI applications, warning that structural changes could invalidate years of development work.
Potential Paths Forward
Mid‑May Trilogue
A tentative next trilogue is scheduled for mid‑May, with the possibility of a simplified agreement that could pave the way for further negotiations.
EU Presidency Transition
Ireland will assume the EU presidency on 30 June, which may influence the direction of negotiations and the final outcome.
Implications for AI Governance Professionals
Regardless of the final agreement, the AI Act remains in force with enforcement of high‑risk systems beginning on 2 Aug 2026. Professionals should prepare for:
- Article 50 compliance, particularly for transparency obligations such as user‑facing disclosures and machine‑readable markings for generative AI.
- Implementation of content‑marking and watermarking capabilities to meet the August 2026 deadline.
- Continuous monitoring of legislative updates, especially concerning sector‑specific carve‑outs and standardisation efforts.
Conclusion
The AI Act Omnibus negotiations illustrate the complex interplay between overarching AI regulation and existing sectoral rules. While a definitive outcome remains uncertain, the high‑risk system deadlines and transparency requirements are set, urging AI governance stakeholders to accelerate compliance preparations.
