EU bans AI nudifier apps and extends deadlines for high-risk AI compliance
What it really says
On May 7, 2026, the European Parliament and the Council of the EU reached a provisional political agreement on amendments to the AI Act. The 'Digital Omnibus' package introduces three key changes: First, AI-powered 'nudifier' applications are explicitly banned - tools that generate non-consensual sexualized deepfake images or videos of real people. The ban also covers AI-generated child sexual abuse material. Violations can be fined up to 35 million euros or 7 percent of annual worldwide turnover, whichever is higher. The prohibition takes effect on December 2, 2026. Second, compliance deadlines for high-risk AI systems were significantly extended: Annex III systems (such as AI in employment, education, or biometric identification) now have until December 2, 2027 instead of August 2026. Annex I systems (AI in regulated products like medical devices) receive an extension to August 2028 instead of August 2027. Third, AI systems in industrial applications already regulated under the Machinery Regulation are exempted from the AI Act. For generative AI systems placed on the market before August 2, 2026, a transitional period for watermarking requirements applies until December 2, 2026. The nudifier ban was partly triggered by the scandal around Elon Musk's AI chatbot Grok, which in early 2026 was able to generate sexualized deepfakes without adequate safeguards.
Our assessment
This agreement is an important step in protecting against one of AI's most tangible dangers: the creation of non-consensual sexualized imagery. Victims are disproportionately women and minors. The EU's explicit action with severe penalties sends a strong signal. At the same time, the deadline extensions for high-risk AI show that industry pressure is having an effect - companies receive up to 16 additional months for compliance. Critics like Substack author Andrew Clearwater argue the EU has 'hit the snooze button on AI regulation.' The truth lies in between: the nudifier ban comes quickly (December 2026), while broader regulation of high-risk systems is postponed. For victims of deepfake abuse, the ban represents concrete progress - provided enforcement works. The challenge remains: nudifier apps can also be operated as open-source software hosted outside the EU.
Relevance for Germany
This agreement is directly relevant for Germany. First, the nudifier ban addresses a growing problem: German schools and police authorities have reported increasing cases of AI-generated nude images of female classmates since 2024. The new ban provides law enforcement with a clear EU-wide legal framework. Second, German companies benefit from the deadline extension: many mid-sized firms had signaled they could not meet the original August 2026 deadline for high-risk AI. The extension to December 2027 provides planning certainty. Third, August 2, 2026 remains the date when the AI Act applies generally - German companies must continue preparing, but with clear prioritization: first the nudifier ban (December 2026), then high-risk compliance (December 2027).
Fact check
Core facts come from the official press releases of the European Parliament and the Council of the EU dated May 7, 2026. The agreement is a 'provisional political deal' - meaning the formal vote is still pending but considered a formality. The deadline extensions (Annex III to December 2027, Annex I to August 2028) and the nudifier ban (from December 2026) are consistently confirmed by multiple legal sources. The penalty of up to 35 million euros or 7 percent of annual turnover corresponds to the AI Act's highest fine tier for prohibited practices. The reference to the Grok scandal is made by Brussels Signal and several other outlets but is not an official justification from EU institutions.
Source
- • European Parliament Press Release 07.05.2026: AI Act: deal on simplification measures, ban on 'nudifier' apps (europarl.europa.eu/news/en/press-room/20260427IPR42011/)
- • Council of the EU Press Release 07.05.2026: Artificial intelligence: Council and Parliament agree to simplify and streamline rules (consilium.europa.eu/en/press/press-releases/2026/05/07/)
- • Brussels Signal 07.05.2026: EU bans AI 'nudifier' apps and delays wider AI Act rules (brusselssignal.eu/2026/05/eu-bans-ai-nudifier-apps-and-delays-wider-ai-act-rules/)
- • Baker Botts / JD Supra 05.2026: EU Parliament and Council of EU Provisionally Agree on Amendments to EU AI Act (jdsupra.com/legalnews/eu-parliament-and-council-of-eu-8989842/)
- • TechPolicy.Press 05.2026: What the EU AI Omnibus Deal Changes for the AI Act and What Lies Ahead (techpolicy.press)