The European Union has decided to postpone the enforcement of several key provisions of its flagship artificial intelligence law, giving technology companies up to two additional years to comply with its strictest requirements. At the same time, the bloc has explicitly banned applications that use AI to generate intimate images without consent.

The decision, approved by the European Parliament on June 16, follows months of pressure from the technology industry, which warned that the technical standards required to implement the law were not yet ready. It follows the political agreement reached on May 7, 2026, between the European Parliament, the Council of the EU, and the European Commission, as part of a broader legislative simplification package known as the “Digital Omnibus on AI.”

The postponement affects compliance deadlines for high-risk AI systems, depending on their category:

Standalone AI systems used in recruitment, employee assessment, lending, education, law enforcement, and critical infrastructure will not be required to comply with the rules until December 2, 2027, instead of August 2026 as originally planned—a delay of 16 months.
AI systems incorporated into products already regulated under separate legislation, such as medical equipment, toys, and radio devices, will have until August 2, 2028, rather than August 2027 as previously scheduled—a delay of 12 months.
Separately, the requirement to label or watermark AI-generated content for systems placed on the market before August 2026 has been postponed from August 2 to December 2, 2026.

European officials insist that this is merely a recalibration of the timetable, not a relaxation of the substance of the law. The stated reason is the absence of harmonized technical standards, which are essential for companies to understand precisely how they must comply.

While the industry welcomed the extension of the deadlines, European lawmakers chose not to compromise on another issue: applications that digitally “undress” real people without their consent will be completely banned from December 2, 2026. Companies developing such tools may face fines of up to €35 million or 7% of their global annual turnover.

Reactions have been divided. The industry welcomed the extended compliance deadlines. At the same time, public-policy researchers warned that the postponement would allow high-risk systems to avoid supervision for well over another year. Digital-rights organizations have also raised concerns about how the ban on “nudify” applications will be enforced, particularly because victims may be required to verify their identity in order to file a complaint, which could make effective protection more difficult.

The decision is not yet final. The package of amendments still requires formal approval by the Council of the EU, followed by publication in the Official Journal of the European Union—a process European officials hope to complete before August 2, when the original rules would otherwise have entered into force.

The AI Act remains the world’s first major law to regulate artificial intelligence according to the level of risk posed by different systems. Adopted in 2024, the law is being applied gradually, and this revised timetable demonstrates how difficult it is, in practice, for regulation to keep pace with a technology that evolves faster than the rules governing it.

Sources:

European Commission — AI Act: Regulatory Framework
European Parliament — AI Act: EP Approves Simplification Measures and “Nudifier” App Ban
Morgan Lewis — EU Approves Delays and Other Amendments to Certain EU AI Act Obligations
Ogletree — EU AI Act Amended: Parliament Votes to Delay Key Deadlines