EU AI Act Omnibus agreed: nudification apps banned, high-risk AI deadlines extended by 16 months
What it really says
On May 7, 2026, the European Parliament and the Council of the EU reached a political agreement on comprehensive amendments to the EU AI Act - the so-called 'AI Act Omnibus', part of a broader digital omnibus package to simplify EU digital regulation. The key changes: First, the use of AI systems to create non-consensual intimate images (nudification apps) and AI-generated child sexual abuse material (CSAM) is banned EU-wide. Violations face fines of up to 35 million euros or 7% of global annual turnover. Second, compliance deadlines for high-risk AI systems under Annex III (biometrics, critical infrastructure, education, employment, law enforcement) are postponed by 16 months - from August 2026 to December 2027. Third, the deadline for transparency requirements for AI-generated content (watermarking) is shortened from 6 to 3 months, with a new deadline of December 2, 2026. Fourth, regulatory exemptions previously available only to SMEs are extended to small mid-cap companies. Formal adoption by Parliament and Council is expected by July 2026.
Our assessment
This agreement is a significant step for European AI protection - and a pragmatic compromise. For those worried about AI-generated deepfakes, the nudification ban is a concrete and overdue safeguard. It shows the EU is willing to respond quickly to new threats. The penalties of up to 35 million euros or 7% of annual turnover are deterrent enough to force even large platforms to act. At the same time, the 16-month deadline extension for high-risk AI should be viewed critically: it gives companies more time for implementation but also means AI systems in sensitive areas like employment and law enforcement can operate longer without the intended safeguards. The shortening of the transparency deadline for AI watermarking from 6 to 3 months is a positive signal: it shows the EU takes labeling of AI content seriously and is accelerating it.
Relevance for Germany
For Germany as the EU's largest economy, these changes are directly relevant. German companies developing or deploying high-risk AI systems - in HR, credit scoring, or medical diagnostics - gain more time for compliance through the December 2027 extension. This affects corporations like SAP, Siemens, and numerous mid-sized companies. The nudification ban closes a protection gap particularly debated in Germany: deepfake pornography disproportionately affects women and minors, and German law enforcement previously had limited tools against the AI systems themselves. The extension of SME exemptions to small mid-caps benefits Germany's economic model, which is strongly shaped by mid-sized enterprises (Mittelstand). Formal adoption is expected before August 2026 - in time before the original high-risk deadlines would have taken effect.
Fact check
The agreement from May 7, 2026 is confirmed by official press releases from the European Commission and the Council of the EU. The specific deadlines - high-risk postponement to December 2027, transparency deadline December 2026 - are consistently reported by the Commission, the Council, and specialist publications including Latham & Watkins, GlobalPolicyWatch, and InsidePrivacy. The penalty framework for the nudification ban (35 million euros / 7%) is documented in the Council press release. Formal adoption by July 2026 is cited as the expected timeframe by all sources.
Source
- • https://ec.europa.eu/commission/presscorner/detail/en/ip_26_1024
- • https://www.consilium.europa.eu/en/press/press-releases/2026/05/07/artificial-intelligence-council-and-parliament-agree-to-simplify-and-streamline-rules/
- • https://ieu-monitoring.com/editorial/eu-reaches-ai-act-omnibus-deal-to-simplify-high-risk-compliance-and-ban-nudification-apps/1193132