MEP Nudifier Deepfakes: EU Ban Enforcement [2026]

On 18 October 2026, the Guardian published the account of Austrian Green MEP Lena Schilling, who learned last month that nudifier deepfakes of her were circulating online. When she opened the links, she found images from one of her first press conferences, digitally altered to remove her clothing. "I felt disgusted," she told the paper. She is one of 50 sitting MEPs, 46 of them women, found on sexualised deepfake sites in a report by the think tank Agora Digitale Transformation (ADT).

The story lands six weeks before the EU's ban on nudifier apps is due to take effect on 2 December 2026. A letter signed by 119 MEPs warns that the ban will fail without enforcement behind it. The case matters well beyond Brussels: the source material for these images was an ordinary, publicly available press-event photo. Anyone who has ever posted a clear picture of their face is, in principle, a target.

This guide walks through what was reported, what the ban does and does not do, what the legal picture looks like, and what practical steps you can take today, including where blurring faces before you publish helps and where it does not.

What Happened to Lena Schilling and the Other MEPs?

ADT, a Berlin-based think tank, found 50 MEPs on deepfake sites; Schilling is one of them. According to the Guardian's report, ADT alerted Schilling to the presence of the images. She said some of the other female MEPs affected had told her they feared speaking out.

EUobserver reported on 5 October that almost 150 European politicians have been targeted by AI-generated deepfake pornography, including those 50 MEPs. According to the Guardian, the earlier ADT report found 138 female parliamentarians from 22 EU countries had been affected, and that women were 33 times more likely than male colleagues to be depicted in or associated with deepfake pornography.

Parliament President Roberta Metsola raised the issue at the opening of the Strasbourg plenary on 5 October. According to Brussels Signal, she said around 60 members, mostly women, had been targeted by websites carrying AI-generated sexual images, from across the political spectrum. She described violence against women, online and offline, as intolerable and deplorable, and said such abuse could not be allowed to become the new normal.

Counts differ: 50 in the ADT report, around 60 in Metsola's remarks. Both are reported by reputable outlets, and the difference is a reminder that nobody has a complete picture. EUobserver quotes the ADT study noting that "the harm is often unseen until it has spread."

The OECD's AI Incidents Monitor has also logged the case as an AI incident rather than a hazard, because the harm has already occurred. Its entry cites 27 articles from outlets in several countries dated 5 to 7 October.

Why Does This Count as Digital Violence and Not Just Online Abuse?

Because the intent and effect are to humiliate and control, and to push women out of public life. Schilling described the images as "digital violence against women" layered on top of the already disproportionate sexist abuse, harassment and violence women in politics face.

She tied it to democracy directly: "These things make going into politics even harder than it is already. Just imagine what would happen to our society if more and more women disappear from public life."

The open letter she initiated with German Green MEP Terry Reintke, as quoted by the Guardian, makes the same argument:

"Its purpose is humiliation and control over women's bodies, and its effect is to make public life more hostile for female politicians. Every young woman considering a political career now must factor in that this could happen to her."

The point for ordinary readers is that the technique does not require sophistication. The letter points to the widespread use of nudifier apps in which photos of women can be undressed with just one click. The victim does not need to have done anything except appear in a photo online.

What Is the EU Nudifier Ban and When Does It Start?

An EU-wide prohibition on AI systems that generate non-consensual sexual imagery takes effect on 2 December 2026. Brussels Signal reports that it was added to the AI Act through the Digital Omnibus package after the Grok affair, and that breaches carry fines of up to 35 million euros or 7 percent of global turnover.

Irish MEP Michael McNamara, who led the file, said the ban targets "nudification apps without consent, which have caused much pain for the profit of some," according to the same report. Compliance deadlines for high-risk AI systems were pushed back to December 2027, and to August 2028 for systems embedded in regulated products, but the nudifier ban was not delayed.

We covered the ban's scope for publishers in our guide to the EU nudifier app ban and consent-safe video publishing. The new development is not the law itself; it is the evidence that harm is happening now and the demand that member states be ready on day one.

Why Do 119 MEPs Say a Ban on Paper Is Not Enough?

Because a prohibition without a regulator, resources and a clear route to act is only a text. The letter, initiated by the Greens and endorsed by MEPs from the Socialists and Democrats, Renew Europe and the Left according to EUobserver, says (as quoted by the Guardian):

"A ban on paper protects no one until it is enforced."

According to EUobserver, the letter asks the European Commission to guide national authorities on enforcing the ban, and asks each member state to have a designated, properly resourced body ready to implement it. Member states are required to set up national authorities to roll out an effective ban from 2 December 2026.

Schilling told the Guardian the goal was to push national authorities to close any loopholes or workarounds that would let the apps keep operating in the EU. Enforcement against a website that can be hosted anywhere is a different problem from enforcement against a store listing, and the letter's concern is that gap between a legal text and a working enforcement route.

There are at least three practical enforcement questions to watch before and after 2 December:

1. Who is the authority in each country? The AI Act relies on national bodies; if one is not named, complaints have nowhere to go.
2. How are offshore sites reached? Blocking, payment-provider pressure and app-store removal are the levers regulators usually have.
3. How fast can victims get content removed? The ban targets the systems; victims still need rapid takedown of the images already made.

Is Sharing a Sexual Deepfake Illegal in the EU?

Increasingly yes, but the timing depends on your country. The separate EU directive on combating violence against women criminalises non-consensual sharing of intimate images, including sexual deepfakes, according to Brussels Signal. Member states have until 14 June 2027 to transpose it into national law. Metsola urged member states to prepare for that transposition and said the ban law was worth no more than the paper it was written on unless enforced.

So there are two different legal tools that people confuse:

The first is aimed at the tool makers; the second at the people who spread the images. A victim needs both to work. Separately, the Commission opened a Digital Services Act investigation into Grok in January, per Brussels Signal, and the Guardian's related coverage notes that xAI is suing Minnesota over a law banning "nudification" technology, a sign that these laws will be tested in court.

For the platform-side duty to label synthetic media, see our explainer on EU AI Act deepfake labeling for creators and platforms. Labeling rules and the nudifier ban are different obligations: a label does not make a non-consensual sexual image acceptable.

Where Do Nudifier Images Come From?

They are made from real photos: yours, your friends', or footage from an event. In Schilling's case the source was a press-conference photo. A nudifier needs a clear, front-facing picture of a person. Public figures have thousands; most of us have fewer, but profile photos, school and club pages, tagged group shots and video stills are plenty.

That has a few practical consequences:

• Video stills count. A single frame from a video is a photo. A clip posted at 1080p or 4K gives a nudifier more usable stills than a thumbnail ever would.
• Other people's posts count. You may be careful, but a tagged group photo from someone else is just as usable. Our guide to unauthorized filming and face blurring covers what you can do when someone else has published you.
• Children are in the same position. Parents who post school events provide the same raw material, which is one reason sharenting has become a safety conversation and not just a privacy one.

The honest limit: you cannot stop a determined person from taking a screenshot of a public photo. What you can do is reduce how much clear, high-resolution, identifiable material you and the people in your care put into public view.

What Can Blurring Do, and What Can't It Do?

Blurring a face before you post removes the part of the image a nudifier most needs to attach to you, but it is not a shield. Honest expectations matter here.

What blurring helps with:

• Event and workplace footage. If you publish video of an event, blurring the faces of people who did not agree to appear removes them as clear subjects. Our guide to AI video blur for privacy protection explains the workflow.
• Children and bystanders. These are the people least able to consent, and the easiest to protect by blurring before upload.
• Reducing the identifiability of public material. A blurred face cannot be tied back to a named person from the image alone.

What it does not do:

• It does not remove images already online. If a photo is already published and scraped, blurring a new copy changes nothing about the old one.
• It does not detect or take down deepfakes. BGBlur is not a deepfake detector or takedown service.
• Weak blurs can fail. Light pixelation can sometimes be partly reversed; see our piece on why weak blur fails. Use a strong blur or solid cover.
• It is not a fit for every case. Politicians, journalists and creators often need their faces visible by definition. For them the answer is layered: privacy settings, watermark-free originals kept off public pages, monitoring, and fast reporting.

BGBlur is a browser-based tool that blurs faces in photos and videos with motion tracking, so a person who moves through a clip stays covered. It also blurs backgrounds and license plates and anonymises voices. It does not claim to prevent misuse, and no blur tool can.

How to Reduce Your Exposure to Nudifier Apps, Step by Step

Start with what you can control: what you publish, where, and at what quality. These steps take under an hour.

1. Audit your public photos. Search your name in an image search and check public profiles, club pages and tagged galleries. Remove or restrict anything you do not need public.
2. Set profiles to private where you can. Friends-only settings reduce scraping from your accounts, though not from other people's posts.
3. Blur faces of people who did not agree to be shown. Before posting event, school or street video, run it through a blur step. With BGBlur's face blur you upload the clip, let the AI detect and track faces, choose the ones to cover and export.
4. Post lower-resolution or cropped versions when high resolution is not needed. This is a mitigation, not a cure.
5. Ask friends and family not to tag or post clear shots of you or your children, and ask for removal when they do.
6. Know your reporting route. Bookmark the platform report forms and your country's police reporting page before you need them.
7. If you are targeted: preserve evidence with URLs and dates, report to the platform and host, file a police report and contact a support organisation. Schilling's message is worth repeating: people misused your photos and you did nothing wrong.

For the legal overview of publishing images of people, see portrait and privacy rights in photography, and for how platform AI features handle your photos, our guide to protecting your photos from Meta's AI image features.

What Should Organisations, Campaigns and Newsrooms Do?

Treat the press-photo pipeline as part of your safety plan. The Schilling case started with a press-conference photograph. Any organisation that publishes images of staff, candidates, volunteers or members should make a few decisions deliberately.

• Decide what resolution and angle go public. Official portraits are useful; a full gallery of clear shots from every event is more than most people need online.
• Blur bystanders and minors by default. Audience members in the background of a speech do not need to be identifiable. Our guide on full body anonymization covers cases where more than the face needs covering.
• Write a response plan. Know who reports, who documents, who supports the person affected, and who talks to the press. Schilling said colleagues feared coming forward; a plan that says "you will be supported" lowers that barrier.
• Monitor. ADT alerted Schilling to the images, which shows that someone has to look. Organisations with high-profile staff can arrange periodic monitoring, with consent.
• Keep the legal timeline in mind. The 2 December ban and the June 2027 transposition deadline give you reference dates for policy updates.

None of this makes harm impossible. It makes it harder to do cheaply, and it makes your response faster when it happens.

What People Are Asking After the Report

Reading how the story spread, three questions come up repeatedly.

Will the ban make the apps disappear? Probably not entirely. Brussels Signal and EUobserver both frame the ban as dependent on national enforcement, and the offshore nature of many sites suggests some will remain reachable. The realistic expectation is fewer mainstream distribution routes, not zero.

Is this only a politician problem? No. ADT's findings concern public figures, but the technique works on any identifiable photo. The Guardian's reporting is specific to MEPs; the underlying risk is general.

Can I trust a tool that says it detects fakes? Detection helps, but it is imperfect, and it is not the same as prevention. Our guide on deepfake fraud statistics and detection tools covers what detectors can and cannot do.

Conclusion

The story of Lena Schilling and the other 49 MEPs is a concrete case study in what the EU is trying to regulate: a press-conference photo turned into a sexualised fake, in a market where, as the open letter says, undressing a woman can take one click. The ban due on 2 December is a serious step, with fines of up to 35 million euros or 7 percent of global turnover reported, and the violence against women directive adds criminal liability for sharing by June 2027.

But as 119 MEPs put it, a ban on paper protects no one until it is enforced. Whether national authorities are named, funded and ready on 2 December is the thing to watch. In the meantime, you control a small but real part of the problem: the clear, high-resolution, identifiable material you publish and allow others to publish.

If you publish video or photos of people, especially children, colleagues or bystanders, blur faces first. BGBlur runs in your browser, tracks faces through a clip and exports a cleaned copy. It will not erase what is already online or stop a determined abuser, but it removes you and the people in your footage as easy source material going forward. If you are targeted, document, report and seek support, and remember that you did nothing wrong.