EU moves to ban AI ‘Nudifier’ apps under expanded Artificial Intelligence Act

European lawmakers are advancing new restrictions on artificial intelligence systems that generate non-consensual explicit content, as part of broader updates to the Artificial Intelligence Act.

Members of the European Parliament have backed a proposal to prohibit so-called “nudifier” apps—AI applications capable of creating or manipulating images and videos to depict individuals in sexually explicit ways without their consent. The measure forms part of a wider amendment package aimed at clarifying timelines and strengthening oversight of high-risk AI systems.

The proposed ban would apply to platforms offering generative AI capabilities, requiring them to implement safeguards that prevent the creation of unauthorized explicit content. Systems that can demonstrate effective protections against such misuse may be exempt from the restriction.

The initiative follows increased scrutiny of AI image-generation features, including controversies surrounding Grok, developed by X. Earlier this year, users identified methods to generate manipulated nude images using the tool, prompting regulatory attention and subsequent limitations by the company. Reports indicated that some workarounds persisted despite those changes.

Regulators are now considering stricter enforcement measures that would apply across the AI ecosystem, including services embedded within major social and content platforms. The rules are expected to introduce clearer accountability for developers and operators of generative AI systems.

In parallel, lawmakers have outlined updated compliance timelines. Requirements for high-risk AI systems—such as those used in infrastructure, law enforcement, and education—are expected to take effect in December 2027, with additional sector-specific provisions following in 2028. Rules mandating watermarking of AI-generated content are scheduled for implementation by November 2026.

The proposal will now move into negotiations between the European Parliament and member states to determine its final form. If adopted, the measures would significantly expand regulatory oversight of generative AI and its potential misuse.