EU Parliament votes for deepfake ban and postpones high-risk AI deadlines to 2027
What it really says
On June 16, 2026, the European Parliament voted 423 to 57 (with 174 abstentions) to approve the 'Digital Omnibus on AI' - the first amendments to the EU AI Act since its adoption in June 2024. The most significant new provision: AI systems that generate or manipulate intimate imagery of identifiable persons without explicit consent ('nudifiers') will be banned from December 2, 2026. AI-generated child sexual abuse material is also explicitly prohibited. Additionally, deadlines for high-risk AI systems were extended: standalone high-risk systems now have until December 2, 2027 (instead of August 2026), while AI systems embedded as safety components in regulated products have until August 2, 2028. Watermarking obligations for AI-generated content were pushed to December 2, 2026. The Council of the EU must formally adopt the amendments before August 2, 2026. Civil liberties organization Liberties.eu criticized the amendments as 'fast-tracking deregulation' and warned of weakened digital rights protections.
Our assessment
This vote merits a green rating because it strengthens citizen protections overall. The explicit ban on AI nudifiers is an important signal: these tools, which can generate realistic nude images of real people with just a few clicks, have proliferated enormously in recent years, disproportionately affecting women and minors. The EU drawing a clear line here is reassuring. The deadline extensions for high-risk AI are a double-edged sword: they give companies more time to prepare, potentially improving implementation quality, but also mean citizens wait longer for full protection. The criticism from Liberties.eu should be taken seriously - the organization warns that some simplifications come at the expense of fundamental rights protection. Overall, however, the positives outweigh the negatives: the nudifier ban arrives quickly (December 2026), and the core structure of the AI Act remains intact.
Relevance for Germany
This vote has immediate implications for Germany. First, the nudifier ban will apply in Germany from December 2026 - a particularly sensitive topic given several recent cases at German schools where AI-generated nude images of classmates were circulated. Second, the extended high-risk AI deadlines give German companies more breathing room. The AI Implementation Act (KI-MIG) passed by the Bundestag on June 11 will need to be adjusted to the new timelines, as the Federal Network Agency is still building up its supervisory capacity. Third, the shifted watermarking deadline (December 2026) aligns with the timeline of the EU Code of Practice on AI-generated content labeling published on June 10, allowing companies to synchronize implementation of both requirements.
Fact check
The vote results (423 yes, 57 no, 174 abstentions) are documented in the official HowTheyVote.eu database. The substantive changes - nudifier ban, extended high-risk AI deadlines (December 2027 and August 2028), and watermarking obligations (December 2026) - are consistently reported by Sofia Globe, Matheson, the EU Council, and Global Policy Watch. The political agreement date (May 7, 2026) is confirmed by the Council's official press release. The criticism from civil liberties organization Liberties.eu is documented in their own analysis published on June 16, 2026.
Source
- • https://howtheyvote.eu/votes/189384
- • https://sofiaglobe.com/2026/06/16/european-parliament-approves-ai-act-amendments-nudifier-ban/
- • https://www.consilium.europa.eu/en/press/press-releases/2026/05/07/artificial-intelligence-council-and-parliament-agree-to-simplify-and-streamline-rules/
- • https://www.liberties.eu/en/stories/ai-omnibus-ep-vote/45715