Tartu: The AI nudifier ban becomes enforceable on 2 December 2026, and developers of image and video tools have less than seven weeks to show that their systems cannot produce non-consensual intimate imagery or child sexual abuse material. The prohibition arrives through the Digital Omnibus on AI, which amends the AI Act and also pushes back the rules for high-risk systems. Legal commentaries report that Parliament approved the text on 16 June and the Council adopted it on 29 June, with entry into force in late July.

The AI nudifier ban adds a new prohibited practice to Article 5 of the AI Act. It targets AI systems that generate or manipulate intimate images of real people without consent, as well as child sexual abuse material, and it reaches the so-called nudifier apps. Providers can rely on a safe harbour if their systems have effective technical safeguards that reliably prevent such outputs.

The reach goes beyond apps built for that purpose. Providers whose systems produce this content as a reasonably foreseeable outcome also fall inside the AI nudifier ban, so developers of general-purpose image and video tools must assess foreseeable misuse in their risk documentation. That shifts the question from intent to design, and it puts technical safeguards at the centre of compliance.

The same omnibus rewrites the calendar for the rest of the law. Obligations for Annex III high-risk systems now apply from 2 December 2027 instead of 2 August 2026, and rules for high-risk AI embedded in regulated products apply from 2 August 2028. National AI regulatory sandboxes have until 2 August 2027. The delay gives providers and regulators more time, although the core duties on data governance, risk management and human oversight stay unchanged.

Beyond the AI nudifier ban, transparency duties move on a separate track. Providers of generative AI must mark synthetic outputs in a machine-readable format under Article 50(2), and that watermarking duty applies from 2 December 2026, with a grace period for systems already on the market before 2 August 2026. The Commission’s code of practice on AI-generated content is still in draft. Its second draft appeared in March 2026, and a final version is expected later this year.

Enforcement of the AI nudifier ban stays tough. The most serious violations still carry penalties of up to €35 million or 7% of global annual turnover, and the AI Act applies to any system that affects people in the EU, wherever the provider is based. The omnibus also widens the AI Office’s supervisory role to AI systems built on general-purpose models from the same provider, with carve-outs for national authorities in areas such as law enforcement and border management.

Compliance teams should treat 2 December as a hard date. Companies that offer image generators in the EU need documented safeguards before that day, and national authorities will look for evidence that foreseeable misuse was assessed. Vendor analyses of the omnibus differ on some details, so providers should check the final text in the Official Journal. The AI nudifier ban will show whether the AI Act can move quickly against a harm that regulators and the public already recognise.