European Parliament delays key EU AI Act requirements and backs ban on nudify apps

The European Parliament voted by a large majority to postpone key deadlines of the EU AI Act — the European Union's flagship law regulating artificial intelligence. At the same time, deputies supported a ban on nudify applications that create fake nude images of real people using AI. The amendments, approved by a significant European Parliament majority, push back the compliance deadlines for developers of high-risk AI systems — those recognized as sources of serious threats to citizens' health, safety, or fundamental rights — until December 2027.

Companies whose AI products are additionally regulated by sector-specific safety standards — for example, in the field of children's toys or medical devices — will have even more time: until August 2028. The EU AI Act was adopted in 2024 and became the world's first comprehensive law regulating AI. The document divides systems into four risk categories: from completely banned (social credit systems, real-time biometric surveillance) to systems with minimal risk.

High-risk systems — algorithms for personnel hiring, medical diagnostics, critical infrastructure management — must undergo independent audits, disclose training data, and comply with strict transparency requirements. The postponement of deadlines was a response to pressure from the technology industry and a number of EU member states, warning that the original timelines would stifle innovation and disadvantage European companies compared to competitors from the US and China. Business representatives repeatedly pointed out: it is impossible to meet all technical and organizational requirements in the originally allotted time — especially for startups without specialized legal departments.

In parallel, parliament voted to ban nudify applications — tools that use AI to create realistic nude images of real people without their consent. Over the past two years, such services have become widespread online: thousands of people, predominantly women, have become victims of this technology — their photos from social networks were turned into compromising content and spread through messengers. The ban on nudify services is part of a broader package of measures against harmful applications of generative AI.

A number of EU countries have already adopted national laws criminalizing certain forms of synthetic content, but a single pan-European standard did not exist until now. The European Parliament's decision closes this gap. The package also included a requirement for mandatory labeling of AI-generated content.

Providers of generative models will be obliged to technically mark the images, videos, and texts they create so that users and platforms can distinguish synthetic content from real — a measure aimed primarily at countering misinformation and manipulation during elections. The adopted amendments must still pass formal approval in the EU Council. Despite the postponements, the basic architecture of the EU AI Act remains unchanged: bans on systems with unacceptable risk have been in effect since February 2025, and rules for general-purpose systems, including large language models, since August 2025.

The European Parliament's decision exposes a contradiction that regulators worldwide face: how to create workable rules for a rapidly developing technology without stifling competitiveness? The postponement gives the industry time to prepare, but human rights advocates warn: the longer oversight of high-risk systems is delayed, the higher the probability of real harm to people who are already today encountering AI in medicine, employment, and banking services.