Laws & Regulations
EU AI Act Nudifier Ban What Changes on 2 December 2026
The European Parliament approved the AI Act "omnibus" with 423 votes in favour. It delays the heaviest compliance deadlines, but it also adds a brand-new prohibition: AI apps that undress real people or generate child sexual abuse material are banned in the EU. Here is what it means for adult platforms, AI generators and companion apps.
423
MEPs voted in favour
2 Dec 2026
Nudifier ban applies
€35M / 7%
Max fine for banned AI
+16 mo
Delay for high-risk AI
01 · The vote
What did the European Parliament approve on the AI Act?
The European Parliament approved a package of amendments to the EU AI Act, known as the Digital Omnibus on AI, on 16 June 2026. It passed with 423 votes in favour, 57 against and 174 abstentions.
The package does two things at once. It postpones and simplifies the most expensive compliance obligations for businesses, and it adds a new banned AI practice targeting nudifier apps and AI-generated child sexual abuse material.
The Council of the EU gave its final approval on 29 June 2026. The amending regulation was published in the Official Journal and entered into force on 27 July 2026, just before the AI Act's general application date of 2 August 2026.
16 Jun 2026
Parliament plenary vote
423 for, 57 against, 174 abstentions.
29 Jun 2026
Council final adoption
Last formal step before publication.
27 Jul 2026
Entry into force
The new deadlines are now legally binding.
What the rapporteurs said
"To all the entrepreneurs and engineers out there, we are pressing the pause button on the AI Act and we are reducing red tape." — Arba Kokalari (EPP, Sweden)
"This omnibus was about establishing legal certainty by extending certain timelines while preserving the AI Act's architecture." — Michael McNamara (Renew, Ireland)
For a wider view of how European and American rules on synthetic sexual content compare, see our breakdown of deepfake and AI porn laws.
02 · The new prohibition
What is the EU nudifier app ban?
The EU nudifier ban is a new prohibited practice under Article 5 of the AI Act. It forbids AI systems that generate non-consensual intimate imagery of real, identifiable people or that generate child sexual abuse material (CSAM).
What exactly is prohibited?
The ban covers AI that creates images, videos or audio showing the intimate parts of an identifiable person, or that person in sexually explicit activity, without their consent. Consent must be freely given, specific, informed and unambiguous.
The ban also covers any AI system that generates or manipulates child sexual abuse material. There is no consent exception for CSAM.
Who does the ban apply to?
The prohibition hits both providers and deployers. Providers are the companies that build or sell the AI tool. Deployers are the businesses that use the tool under their own authority.
A provider breaks the ban when generating this content is the tool's intended purpose, or when it is reasonably foreseeable and reproducible because adequate safeguards are missing.
Is there a safe harbour for adult AI generators?
Yes. A general image or video generator is not banned simply because it can produce explicit content. It stays legal if it has effective technical safeguards that reliably prevent non-consensual intimate imagery of real people and any CSAM.
Fictional, fully synthetic adult characters and content made with the documented consent of the people depicted fall outside the prohibition. The line is drawn at real people who never agreed.
Key insight
The ban targets the function, not the genre. Explicit AI content remains legal in the EU. Undressing a real person without consent does not.
Co-rapporteur Michael McNamara explained the reasoning in plenary: nudifier apps "impact real people, overwhelmingly women, with the purpose of humiliating, degrading and objectifying them."
03 · Timeline
When does the nudifier ban apply?
The EU nudifier ban applies from 2 December 2026. From that date, nudifier and CSAM-generating AI systems can no longer be placed on the EU market or used there.
As of 7 October 2026, that leaves 56 days for providers and deployers to align their systems. The same date is the deadline for machine-readable labelling of AI output from generative systems already on the market.
19 November 2025
Commission proposes the Digital Omnibus on AI
The original package focuses on simplification and delaying high-risk obligations.
7 May 2026
Parliament and Council strike a provisional deal
The final text includes the new ban on nudifier and CSAM-generating AI.
16 June 2026
Parliament plenary approval
423 votes in favour, 57 against, 174 abstentions.
29 June – 27 July 2026
Council adoption and entry into force
The amended deadlines become binding before 2 August 2026.
2 December 2026
Nudifier and CSAM ban applies
Also the labelling deadline for generative AI already on the market.
2 December 2027
Stand-alone high-risk AI obligations apply
Covers Annex III uses such as biometrics, hiring and education.
2 August 2028
Embedded high-risk AI obligations apply
Covers AI used as a safety component in regulated products.
Days left until each AI Act deadline (from 7 October 2026)
Source: Inside Intelligence Unit calculation on dates set by the Digital Omnibus on AI.
04 · Postponements
Which AI Act deadlines were delayed?
The omnibus delays obligations for high-risk AI systems by 12 to 16 months and gives existing generative AI tools four extra months to add machine-readable labels.
Stand-alone high-risk systems, such as biometric identification or AI used in hiring, now have until 2 December 2027. AI built into regulated products like machinery or medical devices has until 2 August 2028.
The nudifier ban runs in the opposite direction. It is a new obligation with a short runway, not a postponement.
| Obligation | Original date | New date | Change |
|---|---|---|---|
| Nudifier & CSAM ban (Art. 5) | Not in the original Act | 2 Dec 2026 | New prohibition |
| Labelling of AI output, existing tools (Art. 50(2)) | 2 Aug 2026 | 2 Dec 2026 | +4 months |
| National AI regulatory sandboxes | 2 Aug 2026 | 2 Aug 2027 | +12 months |
| Stand-alone high-risk AI (Annex III) | 2 Aug 2026 | 2 Dec 2027 | +16 months |
| Embedded high-risk AI (Annex I) | 2 Aug 2027 | 2 Aug 2028 | +12 months |
Extra months granted by the AI omnibus
Source: European Parliament; Digital Omnibus on AI.
What else did the omnibus simplify?
- Small mid-cap companies now get the same lighter documentation and capped fines as SMEs.
- Providers may process sensitive personal data strictly to detect and correct bias, with safeguards.
- Overlapping requirements with the EU machinery rules were removed.
- Tools that only optimise performance no longer count as a "safety component".
- Supervision of AI systems built on general-purpose models is centralised at the EU AI Office.
05 · Transparency
Do adult sites have to label AI-generated content?
Yes. Under Article 50 of the AI Act, providers of generative AI must mark synthetic images, video and audio in a machine-readable format, and deployers must disclose deepfakes to viewers.
Generative tools placed on the EU market after 2 August 2026 must comply immediately. Tools already on the market before that date have until 2 December 2026 to add watermarks or metadata.
For deployers, any AI-generated or manipulated content that resembles real people, places or events counts as a deepfake. It has to be clearly disclosed as artificial to the people who see it.
What does this mean for AI adult content?
An AI companion app or image generator serving EU users needs output marking built into the generation pipeline, not added by hand. Metadata standards and invisible watermarks are the usual routes.
Sites publishing AI scenes of fictional characters should still label them clearly. Our analysis of the AI-generated adult content market shows how fast this segment is growing, which is exactly why regulators are paying attention.
Teams building their own generation stack can check which models and fine-tunes they rely on in our guides to AI models for adult content generation and adult LoRA models.
06 · Industry impact
How does the nudifier ban affect the adult industry?
The nudifier ban mainly affects AI image and video tools that accept photos of real people. Platforms built on consenting performers or fully fictional AI characters face labelling duties, not a ban.
| Business type | Exposure | What to check |
|---|---|---|
| Apps that "undress" uploaded photos | Banned in the EU | The core function is the prohibited practice. No safeguard fixes it. |
| NSFW image & video generators | High | Block real-person likeness and photo uploads used for sexual edits; robust CSAM filters. |
| AI companion apps with images | Medium | Fictional characters only; no "make her look like" features from user photos; label outputs. |
| Tubes & UGC platforms hosting AI clips | Medium | Deepfake disclosure, fast takedown of non-consensual content, uploader consent records. |
| Directories & review sites | Indirect | Do not list or promote tools whose selling point is undressing real people. |
| Creators using AI on their own content | Low | Label AI-edited content; keep written consent from every person shown. |
Why are image-upload features the biggest risk?
A tool that lets users upload a photo of a real person and request a sexual edit makes non-consensual output reasonably foreseeable and reproducible. That is the exact test the new Article 5 uses for providers.
Removing or tightly gating these features is the fastest way to move a product outside the ban. Keyword blocklists alone are unlikely to count as effective safeguards.
What about AI companion apps?
AI companion apps that generate images of original, fictional characters are not the target of the ban. Their main EU duty is labelling synthetic output under Article 50.
Builders can review safe architecture choices in our AI companion app architecture guide and the AI companion tech stack.
Does the EU ban affect non-EU adult platforms?
Yes. The AI Act applies to any provider whose system is placed on the EU market or whose output is used in the EU, wherever the company is based.
US-based operators facing state-level rules at the same time can compare obligations in our US legal compliance guide for AI adult sites.
Content moderation will carry most of the load. Our report on AI content moderation and false positives explains why automated filters need human review behind them.
07 · Enforcement
What are the fines for breaking the nudifier ban?
Breaking a prohibited-practice rule under the AI Act can cost up to €35 million or 7% of worldwide annual turnover, whichever is higher. The nudifier ban sits in this top tier.
The omnibus left the penalty structure of the AI Act unchanged. For SMEs and small mid-caps, the lower of the two amounts applies.
National market surveillance authorities enforce the ban. Victims of non-consensual imagery can also rely on national criminal law and the Digital Services Act for takedowns.
AI Act maximum fines as % of global annual turnover
Source: Regulation (EU) 2024/1689 (AI Act), Article 99. Higher of the two amounts applies, lower for SMEs.
08 · Political support
How did MEPs vote on the AI omnibus?
The AI omnibus passed with 64.7% of votes cast in favour. Only 8.7% voted against, while 26.6% abstained.
The high abstention rate reflects a split inside Parliament. Some MEPs backed the nudifier ban but opposed delaying protections against high-risk AI.
The ban on nudifier apps itself was one of the least contested parts of the package. It was the main addition Parliament secured during negotiations.
European Parliament vote, 16 June 2026 (654 votes cast)
Source: European Parliament press room.
09 · Action plan
What should adult platforms do before 2 December 2026?
Adult platforms using generative AI should audit every feature that touches real people's images, add machine-readable labels to AI output and document their safeguards before 2 December 2026.
Pre-deadline compliance checklist
- Map every AI feature that generates or edits images, video or audio.
- Remove or gate photo uploads that can be used to sexualise a real, identifiable person.
- Deploy CSAM detection on both prompts and outputs, with human escalation.
- Collect and store written consent for every real person in AI-assisted content.
- Add machine-readable marking (metadata or watermark) to all synthetic output.
- Show a visible "AI-generated" label wherever deepfake-style content is published.
- Review partners and affiliate links: drop any tool marketed for undressing real people.
- Write down your safeguards. Documentation is your first defence in an audit.
Key insight
Consent records do double duty. They protect against the nudifier ban and they support privacy protection for adult creators.
The nudifier ban adds to a stack of rules already reshaping adult traffic in Europe. Our study on how age verification affects porn search traffic shows what happens when compliance rules hit real users. More analysis lives in our Laws & Regulations hub.
Key takeaways
- The EU AI omnibus passed Parliament on 16 June 2026 with 423 votes in favour and entered into force on 27 July 2026.
- Nudifier apps and CSAM-generating AI are banned in the EU from 2 December 2026.
- Explicit AI content stays legal. The ban targets real people depicted without consent.
- Generators stay legal only with effective safeguards that reliably prevent prohibited output.
- High-risk AI obligations moved to December 2027 and August 2028.
- Fines reach €35 million or 7% of global turnover for prohibited practices.
