European Union lawmakers have reached a provisional agreement to soften key timelines in the bloc's sweeping AI Act, granting enterprises additional time to prepare for high-risk compliance obligations. The deal, struck between negotiators for the European Parliament and the European Council early Thursday, pushes back the strictest deadlines and clarifies overlapping rules, aiming to reduce administrative burdens on businesses.
Key Changes to Deadlines
Under the provisional agreement, high-risk AI systems will now face revised compliance deadlines. Stand-alone high-risk systems must meet requirements by December 2, 2027, while AI used in products covered by EU sectoral safety rules — such as medical devices, toys, lifts, machinery, and watercraft — must comply by August 2, 2028. This represents a significant shift from the original uniform deadline of August 2, 2026, as outlined in a European Parliament statement.
The delay gives businesses nearly two extra years to align with the AI Act's strict rules for systems that pose potential risks to health, safety, or fundamental rights. However, the deal still requires formal adoption by both the Parliament and the Council before it can become law. The co-legislators aim to complete this step before August 2, 2026; until then, the original deadline remains in effect as drafted.
Regulatory Sandboxes and Watermarking Obligations
Member states now have an extra year to set up AI regulatory sandboxes — controlled environments where companies can test AI systems under regulatory supervision. The deadline for establishing these sandboxes has been pushed back to August 2, 2027, according to the Council's statement.
Conversely, watermarking obligations on AI-generated content will be implemented earlier than the European Commission originally proposed. The new deadline for these transparency measures is December 2, 2026, rather than February 2, 2027, as noted by the Parliament.
Reducing Overlap and Clarifying Rules
The provisional agreement aims to streamline regulatory requirements by eliminating overlapping rules for AI integrated into machinery products. Under the new terms, such AI systems will follow only sectoral safety rules, with safeguards ensuring equivalent health and safety protection. This prevents dual compliance under both the AI Act and existing machinery directives.
Furthermore, the definition of a "safety component" under the AI Act has been narrowed. AI features that merely assist users or improve performance — without posing health or safety risks in case of failure — will no longer be automatically treated as high-risk. The Parliament emphasized that this change targets systems where a malfunction does not create direct dangers.
For broader sectors like medical devices, toys, lifts, machinery, and watercraft, the co-legislators agreed on a mechanism to resolve overlaps between the AI Act and existing sectoral laws. This aims to provide clarity and reduce compliance complexity for manufacturers already subject to stringent industry-specific regulations.
Impact on Businesses
Small and medium-sized enterprises (SMEs) are not the only beneficiaries. The provisional deal extends exemptions previously limited to SMEs to small mid-cap companies — those with fewer than 500 employees. This expansion gives more midsize firms breathing room to adapt to the AI Act's requirements.
The agreement also clarifies the division of supervisory responsibilities. The EU AI Office will centrally oversee general-purpose AI systems, ensuring consistent enforcement across the bloc. Meanwhile, national authorities will retain responsibility in areas such as law enforcement, border management, judicial authorities, and financial institutions. This division aims to balance centralized oversight with local expertise.
Next Steps and Political Context
The breakthrough comes just nine days after previous discussions collapsed without agreement. The Cypriot presidency of the Council played a key role in brokering the deal. Marilena Raouna, Cyprus's deputy minister for European affairs, stated: "Today's agreement on the AI Act significantly supports our companies by reducing recurring administrative costs."
Arba Kokalari, the Parliament's co-rapporteur for the Internal Market and Consumer Protection committee, added: "With this agreement, we show that politics can move just as quickly as technology." The deal now heads to formal adoption by the full Parliament and Council. If approved, it will provide much-needed clarity and flexibility for businesses navigating the EU's landmark AI regulation.