EU Reaches Provisional Deal to Amend and Delay Parts of Landmark AI Act

Negotiators from the European Parliament and Council reached a provisional agreement on the AI Omnibus at 4 am on Thursday. The AI Omnibus is an EU effort to simplify AI rules after months of heated debates over exemptions, legal challenges and concerns that the package could undermine the original AI Act.

The deal limits removal of overlapping AI Act provisions with sectoral rules to machinery products only.

Overlaps for medical devices, toys, lifts and watercraft will instead be addressed through implementing acts. ” Personal data processing will be permitted to detect and correct biases in both high-risk and non-high-risk AI systems. SME exemptions were extended to small and mid-caps with up to €200 million euro turnover.

Euronews reported that the change is intended to support EU tech scale-ups. Stricter rules will apply to AI systems that can be used to create child sexual abuse material or non-consensual deepfake nudity and sexually explicit content. Those systems must comply by December 2 2026.

The application of high-risk rules was set for 2 December 2027 for stand-alone high-risk AI systems. For high-risk AI systems embedded in products the date is 2 August 2028. The grace period for watermarking AI-generated content was set at 2 December.

The deadline for AI regulatory sandboxes was set for 2 August 2027. That represents a one-year delay from the original planned deadline of August 2 2026. Euronews reported the extension reflects broader enforcement delays.

The European Parliament and Council must still formally approve the AI Omnibus. While key negotiators celebrated the outcome, doubts remain about whether the measures will sufficiently help European businesses navigate complex rules few appear ready to meet. Attention has now shifted to the Digital Omnibus package focused on AI development and data.

The Central European AI Chamber and 15 other associations, including the Consumer Choice Center Europe, shared an open letter urging EU Member States to correct course on watered-down Digital Omnibus proposals. The open letter calls for a better balance between data protection and the EU’s broader strategic goals such as innovation and economic growth.

Key pain points include the definition of personal data, AI-related processing exemptions under legitimate interest basis (Article 88c) and the incidental processing of sensitive data (Article 9(5)).

Current proposals for the scope of scientific data for AI development are being narrowed. Under the Digital Omnibus proposals the “cookie fatigue” mechanism may be replaced by a new mechanism under Article 88b. The signatories argue that the Commission’s initial proposals on pseudonymization and AI-related exemptions were pragmatic while current drafts return toward the status quo.

Without clear distinctions companies and researchers face fragmented interpretations across Member States and must rely on guidelines not designed to promote economic growth. This story was originally published on EU Tech Loop and has been shared on Euronews as part of an agreement.