The European Union has reached a provisional deal to amend the AI Act, pushing the enforcement deadline for high-risk AI system regulations by 16 months to December 2027. Originally entering into force in August 2024, the legislation covers AI use in biometrics, critical infrastructure, law enforcement, and employment. The delay aims to provide regulatory relief for businesses facing competition from China and the United States, though critics argue it favors large technology companies over fundamental rights.

Alongside the delay, the agreement introduces an immediate and total prohibition on 'nudifier' applications, which are AI tools capable of generating non-consensual explicit images, video, and audio of individuals. This provision explicitly includes AI-generated child sexual abuse material and is expected to take effect on December 2, 2026. The move follows reports in early May 2026 of AI-generated intimate content depicting Italian Prime Minister Giorgia Meloni.

Organizations violating this new ban face financial penalties of up to €35 million or 7% of their total global annual turnover. While the current amendments await final formal approval, the AI Act continues to be classified as the world's strictest regulatory framework for artificial intelligence. The updated timeline allows for a more phased implementation of compliance requirements for developers and operators working within the EU member states.

The European Union has reached a provisional deal to amend the AI Act, pushing the enforcement deadline for high-risk AI system regulations by 16 months to December 2027. Originally entering into force in August 2024, the legislation covers AI use in biometrics, critical infrastructure, law enforcement, and employment. The delay aims to provide regulatory relief for businesses facing competition from China and the United States, though critics argue it favors large technology companies over fundamental rights.

Alongside the delay, the agreement introduces an immediate and total prohibition on 'nudifier' applications, which are AI tools capable of generating non-consensual explicit images, video, and audio of individuals. This provision explicitly includes AI-generated child sexual abuse material and is expected to take effect on December 2, 2026. The move follows reports in early May 2026 of AI-generated intimate content depicting Italian Prime Minister Giorgia Meloni.

Organizations violating this new ban face financial penalties of up to €35 million or 7% of their total global annual turnover. While the current amendments await final formal approval, the AI Act continues to be classified as the world's strictest regulatory framework for artificial intelligence. The updated timeline allows for a more phased implementation of compliance requirements for developers and operators working within the EU member states.