Nudifier Ban — New Prohibition in Article 5 of the EU AI Act

What is the nudifier ban?

The Omnibus VII legislative package introduced a new prohibited practice under Article 5(1)(i) of the EU AI Act. It prohibits AI systems that generate synthetic sexually explicit images or videos of real persons without their explicit consent.

This provision directly targets so-called "nudifier" or "undressing" AI tools that use generative models to remove clothing from photographs of persons, as well as AI tools for generating deepfake pornographic content.

Text of the provision

Article 5(1)(i) prohibits:

The placing on the market, putting into service or use of AI systems that generate synthetic sexually explicit images or videos of persons using biometric data or the likeness of real persons, without the explicit, specific and informed consent of those persons.

What the prohibition covers

Prohibited activities

1. Nudifier/undressing AI — tools that "undress" persons in photographs using generative models
2. Deepfake pornography — generation of pornographic content featuring the face/body of a real person
3. CSAM generation — generation of child sexual abuse material (absolute prohibition with no exceptions)
4. Distribution — placing such tools on the market within the EU

Who is covered

The prohibition applies to:

• Providers — those who develop or place the system on the market
• Deployers — those who use the system within the EU
• Distributors — those who enable access to the system

Exceptions

The prohibition does not cover:

1. Consent — If the person whose likeness is used has given explicit, specific and informed consent, the system is not prohibited
2. Fictional characters — Generation of content that does not use the likeness of real persons is not covered by Art. 5(1)(i), but may be subject to Art. 50 (transparency)
3. Law enforcement — Use in the context of criminal prosecution under strict judicial oversight

Note: For minors, there is no consent exception — the prohibition is absolute.

Penalties

Violation of Article 5 carries the highest level of penalties under the EU AI Act:

Penalty	Amount
Maximum fine	EUR 35,000,000 or 7% of total annual global turnover, whichever is greater
For SMEs/startups	Proportionately lower, but still significant

Why this was added

Omnibus VII introduced this prohibition in response to:

• Growing problem — Europol and national agencies reported an explosion of nudifier tools in 2025
• Gender-based violence — The majority of victims are women and girls; the EP recognised this as a form of gender-based violence
• Child protection — AI-generated CSAM became a growing problem requiring an explicit prohibition
• Regulatory gap — The previous text of Art. 5 did not explicitly cover this type of system

Practical implications

For ComplianceForge users

If you are using the ComplianceForge wizard, Step 4 (Prohibited Practices) automatically checks whether your organisation uses AI systems that could fall under this prohibition.

For AI system developers

If you develop generative AI systems that work with images or video:

1. Verify whether your system can generate sexually explicit content
2. Implement technical safeguards (content filters, safety classifiers)
3. Document security measures in your technical documentation
4. Monitor usage to detect misuse

Application date

The prohibition takes effect immediately upon publication of Omnibus VII in the Official Journal of the EU — there is no transitional period for Art. 5 provisions.

Related articles

• Guidelines on Prohibited AI Practices — Original prohibitions under Art. 5
• Omnibus VII — What Changed — Overview of all amendments

Sources

• EU Parliament — Omnibus Simplification Regulation
• AI Act Art. 5 — Prohibited Practices
• Europol — Threat Assessment on AI-generated CSAM