The European Parliament approves amendments in the AI Act and bans nudification applications with artificial intelligence.

The European Parliament has approved amendments to the AI Act that postpone certain obligations for high-risk AI systems, reduce overlaps with regulations for machinery products, and prohibit applications for nudification using artificial intelligence. The text is part of the digital omnibus package and still needs to be formally approved by the Council of the European Union before it comes into force.

The European Parliament has approved amendments to the AI Act that postpone the application of certain obligations for high-risk AI systems, reduce overlaps with legislation regarding machinery products, and prohibit applications for nudification using artificial intelligence. The vote took place in plenary and concluded with 423 votes in favor, 57 against, and 174 abstentions.

In short

1. The European Parliament has approved amendments to the AI Act as part of the digital omnibus package.

2. Obligations for high-risk autonomous AI systems will apply from December 2, 2027.

3. Obligations for AI systems integrated as safety elements in sector-regulated products will apply from August 2, 2028.

4. Watermarking obligations for AI-generated content are postponed until December 2, 2026.

5. The text prohibits AI systems that generate child sexual abuse material or explicit or intimate images, videos, and audio of an identifiable person without their consent.

The approved amendments aim to support companies in implementing the AI Act while maintaining the core architecture of the legislation and a risk-based approach. The European Parliament presents the text as part of the digital omnibus package proposed by the European Commission to simplify certain digital rules.

The new deadlines postpone the application of certain obligations for high-risk AI systems so that the necessary standards and support measures are available before the rules become applicable. For high-risk autonomous AI systems, obligations will start on December 2, 2027.

For AI systems integrated as safety elements and covered by the Union's sectoral legislation on safety and market surveillance, the deadline is August 2, 2028. At the same time, most provisions of the AI Act will start to apply from August 2, 2026.

The text also postpones watermarking obligations for AI-generated content until December 2, 2026. Until then, AI-generated content will need to be labeled in a way that can be automatically read to increase transparency.

One of the most important amendments is the prohibition of AI systems that can generate child sexual abuse material or that can create images, videos, and audio of intimate parts or explicit sexual activities of an identifiable person without their consent.

Providers will not be able to introduce such systems into the European Union market, except in cases where they include adequate technical safeguards to prevent the creation of such material. The prohibition also applies to users who use the systems for this purpose.

Companies will have until December 2, 2026, to bring their systems into compliance with the new requirements. The Parliament presented this prohibition as a strengthening of the protection of individuals affected by abusive uses of artificial intelligence.

Arba Kokalari, co-rapporteur for the Internal Market and Consumer Protection Committee, said in the plenary debate that the text aims to simplify the application of the AI Act for firms and reduce bureaucracy.

"For all the entrepreneurs and engineers out there, we are hitting the pause button on the AI Act and reducing bureaucracy," Kokalari said. She added that it should be easier to build the technology companies of the future in Europe and to stay in Europe so that the Union can become "an AI continent."

The MEP said that the text eliminates the duplication of rules for machinery products and reduces reporting obligations for companies. She also mentioned the prohibition of AI systems that generate child sexual abuse material.

Michael McNamara, co-rapporteur for the Civil Liberties, Justice, and Home Affairs Committee, presented the file as a compromise between legal clarity, maintaining the architecture of the AI Act, and strengthening protections where possible.

"This omnibus was about establishing legal certainty by extending certain deadlines while maintaining the architecture of the AI Act and strengthening protections where possible," McNamara said.

He said that the Parliament achieved a complete ban on applications for nudification using artificial intelligence. McNamara described these tools as having real effects on people, especially women, being used for humiliation, degradation, and objectification.

"I am proud that this Parliament fought for the ban, which will come into force before the end of this year," said the Irish MEP.

Other amendments clarify the relationship between the AI Act and legislation regarding machinery products. Overlapping requirements for AI used in such products are removed, and products will need to comply with sectoral safety legislation, maintaining an equivalent level of health and safety protection.

The text also clarifies what "safety element" means. Products with AI functions that only assist users or optimize performance will not automatically fall under high-risk obligations if their failure or malfunction does not create risks to health or safety.

The amendments also allow for the processing of personal data where strictly necessary for detecting and correcting biases, with appropriate safeguards, both in high-risk AI systems and in those that are not high-risk. The text extends some exemptions applicable to SMEs and to medium-capitalized small enterprises to support their growth.

The application of certain rules for general-purpose AI systems will be simplified through a strengthened role of the EU's AI Office. The Parliament presents this centralization as part of an effort to avoid institutional overlaps and divergent interpretations.

The text still needs to be formally approved by the Council of the European Union before it comes into force.

The approved legislation is part of the eighth omnibus simplification package proposed by the European Commission on November 19, 2025. The European Parliament is also working in parallel on two other proposals from the digital omnibus package, one regarding amendments to legislation on data use and data protection, and the other regarding European wallets for businesses.

https://2eu.brussels/ro/stiri/parlamentul-european-aproba-amanari-in-ai-act-si-interzice-aplicatiile-de-nudificare-cu-inteligenta-artificiala