The European Parliament supports the postponement of key rules from the AI Act and proposes a ban on "nudifier" applications.

The IMCO and LIBE committees of the European Parliament support the postponement of the application of essential requirements for high-risk artificial intelligence systems, citing the lack of necessary technical standards until the current deadline of August 2026. At the same time, MEPs propose a ban on AI applications that generate sexual images without consent and seek to reduce overlaps between the AI Act and EU sectoral legislation.

The European Parliament supports the postponement of important provisions of the AI Act, especially for high-risk artificial intelligence systems, namely those applications used in sensitive areas such as critical infrastructure, education, labor market, or law enforcement. In parallel, MEPs propose more flexibility for companies and introduce an explicit ban on "nudifier" applications, which use AI to create sexual images of real people without their consent.

In short

The IMCO and LIBE committees adopted the common position with 101 votes in favor, 9 against, and 8 abstentions.

The Parliament supports the postponement of rules for high-risk AI due to delays in technical standards.

The new proposed deadlines are December 2027 and August 2028.

MEPs propose a ban on "nudifier" AI applications.

Measures for flexibility and reduction of legislative overlaps are introduced.

The European Parliament has adopted its position on the proposal to simplify the AI Act, supporting the postponement of obligations for high-risk artificial intelligence systems. The central argument is that the necessary technical standards for the application of these rules may not be finalized by the current deadline of August 2, 2026.

To avoid legal uncertainty, MEPs propose clear application deadlines. Thus, high-risk AI systems explicitly listed in the regulation, including those used in areas such as biometrics, critical infrastructure, education, or justice, would be subject to the rules starting from December 2, 2027. In the case of AI systems integrated into products already regulated by other European safety laws, the proposed deadline is August 2, 2028.

The Parliament also supports granting an additional deadline for marking AI-generated content, known as "watermarking," that is, labeling audio, video, or text content to indicate that it is artificially created. However, MEPs propose a shorter postponement than that advanced by the Commission, until November 2, 2026.

One of the most visible changes is the introduction of a ban on "nudifier" applications, that is, AI-based tools that can generate or manipulate images to create explicit sexual content using the face or body of a real person without their consent. The ban would not apply to systems that include effective mechanisms to prevent such uses.

The text also reflects pressure from the industry for more flexible rules. MEPs argue that providers should be able to process personal data to detect and correct biases in AI systems, but only when strictly necessary and with adequate safeguards.

At the same time, the Parliament supports extending support measures to small mid-cap enterprises, that is, firms larger than SMEs but not as large as major corporations, to facilitate their growth. It also seeks to reduce overlaps between the AI Act and sectoral legislation, arguing that obligations can be less strict in the case of products already regulated by other European norms.

The rapporteur for the internal market, Arba Kokalari, explained that the goal is "predictable, stop-the-clock rules, simplified to eliminate overlaps with sectoral legislation and reduce fragmentation among member states," stating that such "stop-the-clock" mechanisms mean temporarily suspending certain deadlines or obligations to give companies more time to comply. She added that "companies now need clarity on whether they are high risk or not" and that "if Europe wants to be competitive, we need to increase investments and make AI easier to use, not punish companies that introduce innovative AI features in safe products."

The rapporteur for civil liberties, Michael McNamara, emphasized that a broad compromise was possible and welcomed the inclusion of the ban on "nudifier" applications, stating that this is "something our citizens expect from co-legislators."

The next step is the plenary vote, expected on March 26, after which negotiations with the Council can begin.

The proposal is part of the digital simplification package promoted by the European Commission and shows that EU institutions are trying to adjust the implementation of the AI Act before some of the most complex obligations come into force. The lack of technical standards and the risk of legal uncertainty are the main reasons cited for the postponement.

At the same time, the adopted compromises show that the Parliament is trying to maintain a balance between supporting innovation and protecting users. The postponement of certain rules and the flexibility of the framework are accompanied by the introduction of specific bans, especially for applications considered high risk for fundamental rights.

https://2eu.brussels/ro/stiri/parlamentul-european-sustine-amanarea-unor-reguli-cheie-din-ai-act-si-propune-interzicerea-aplicatiilor-de-tip-nudifier