Full Breakdown
EU AI Act Reform Stalls as Parliament and Council Fail to Reach Deal
4/30/2026, 6:56:04 AM
Failed AI Act Reform Talks
On 28 April 2026, EU member-state representatives and the European Parliament held a 12-hour trilogue in Brussels to amend the AI Act within the Digital Omnibus. Negotiations began at 1100 GMT and collapsed early on 29 April, leaving the reforms unresolved.
Background & Digital Omnibus
The AI Act, effective August 2024, imposes strict rules on high-risk AI, covering biometric ID, utilities, health, credit scoring and law-enforcement. The Digital Omnibus, introduced November 2025, aims to align the AI Act with the GDPR, e-Privacy Directive and Data Act, and had previously agreed to postpone compliance for standalone high-risk systems to 2 December 2027 and for AI embedded in regulated products to 2 August 2028.
Timeline of the Negotiations
28 April 2026 1100 GMT – trilogue opens; 28 April ~2300 GMT – talks stall over sector-specific carve-outs; early 29 April – negotiations adjourned. Sources indicate a follow-up session is expected in roughly two weeks, though no exact date has been set.
Official Statements & Responses
Cyprus, holding the rotating Council presidency, said an agreement was impossible. German Chancellor Friedrich Merz and MEP Svenja Hahn pushed for exemptions for industrial AI. Dutch Green MEP Kim van Sparrentak warned of regulatory chaos. EU Commission Vice-President Henna Virkkunen called for a swift, high-quality text. EPP co-rapporteur Arba Kokalari urged more time to simplify the law.
Deadlines & Compliance Numbers
The original high-risk AI compliance date is 2 August 2026. Proposed postponements would shift standalone high-risk systems to 2 December 2027 (Annex III) and AI embedded in regulated products to 2 August 2028 (Annex I). The talks lasted roughly 12 hours.
Impact on Industry and Competition
Delaying or weakening the AI Act could reduce compliance costs for manufacturers of machinery, medical devices and other high-risk products, but critics argue it may create a legal vacuum that harms European competitiveness against U.S. and Asian firms. The outcome also affects data-privacy and cybersecurity safeguards.
Opposition and Uncertainties
Green MEP van Sparrentak accused the European People’s Party of risking a “total legal vacuum of chaos.” Over 40 civil-society groups warned that sector-specific carve-outs could lead to deregulation. Sources differ on the timing of the next trilogue and on whether the Council will accept the Parliament’s push to move product-embedded AI to sectoral law.
Verbatim Quotes
- “Big Tech is probably popping champagne. While European companies that care about safety and did their homework now face regulatory chaos,” — Kim van Sparrentak, Dutch MEP
- “Parliament’s proposal is clear: remove industrial AI from the AI Act and apply sector-specific laws to end double regulation,” — Svenja Hahn, German liberal lawmaker
- “The breakdown in talks on the EU’s AI rules is a setback,” — Ben Brake, DOT Europe director general
- “It was not possible to reach an agreement with the European Parliament,” — Cypriot official, EU Council presidency
What’s Next
The next trilogue is slated for mid-May, with the Irish presidency taking over on 30 June. If no deal is reached before the 2 August 2026 deadline, the original high-risk obligations will apply as written.
