The recent negotiations surrounding the European Commission's AI Act Omnibus highlight significant challenges and implications for AI governance in the EU. After extensive discussions, the trilogue on April 28 failed to reach an agreement, raising questions about the timeline and enforcement of regulations for high-risk AI systems.
The negotiations, aimed at finalizing the Omnibus package introduced in November 2025, revealed deep divisions among the European Parliament, Council, and Commission. A central issue was whether AI systems integrated into products governed by existing EU safety legislation should be exempt from additional AI Act requirements. While the Parliament advocated for such exemptions, the Council and Commission opposed them, emphasizing the need to maintain the AI Act's overarching regulatory framework.
One of the key sticking points involves Annex I, which outlines products covered by EU safety regulations. The Parliament's push to shift certain high-risk AI systems out of the AI Act's direct oversight could lead to a deregulatory approach, potentially undermining the intended protections. Stakeholders, including civil society organizations, have expressed concern over this potential shift.
Looking ahead, the next trilogue is scheduled for mid-May, with the possibility of new dynamics emerging as Ireland assumes the EU presidency at the end of June. Despite the setbacks, the enforcement deadline for high-risk systems remains set for August 2026, and stakeholders are urged to prepare for upcoming transparency requirements for generative AI systems. The outcome of these negotiations will be crucial in shaping the future landscape of AI governance in Europe.