At least 138 women members of parliament across 22 European Union countries have appeared on or been referenced by sexually explicit deepfake websites, according to research from German think tank Agora Digitale Transformation. Only 9 male MPs were referenced across the same ecosystem — a 33-to-1 gender skew that the researcher behind the work, Benjamin Shultz, calls a direct measure of how targeted the abuse has become.
The analysis covered roughly 160 domains hosting deepfake abuse or other nonconsensual sexual content. Shultz, a fellow at Agora and lead researcher at the American Sunlight Project, compiled records on more than 5,800 members of parliament across the EU's 27 countries in July, then used custom search tooling to find where those names, faces, and profiles surfaced across the deepfake network. Members of the European Parliament were excluded from the sweep.
Politicians from Germany, the Netherlands, Italy, and France showed up most often, and seniority correlated with exposure — the more senior the woman, the more likely her likeness had been ingested. Suzanne Kröger, a Dutch MP with Progressive Netherlands who was among those affected, called the findings deeply disturbing.
“Deepfake porn is part of a much bigger trend of online hate and intimidation against women and female politicians”— Suzanne Kröger, Dutch MP, Progressive Netherlands
Key facts
- 01138 women MPs across 22 EU countries appeared on or were referenced by deepfake abuse sites, versus 9 male MPs.
- 02Researcher Benjamin Shultz analyzed roughly 160 domains and cross-referenced them against 5,800 European parliamentarians compiled in July.
- 03Women MPs are 33 times more likely than men to be targeted, with Germany, the Netherlands, Italy, and France seeing the highest counts.
- 04At least half a dozen of the domains identified in the research have been removed from the web ahead of publication.
- 05The UK and EU are drafting bans on nudify services; the US Take It Down Act has already forced two major deepfake sites offline.
Kröger noted that the visibility of elected officials makes the pattern legible in a way it isn't for ordinary victims. "We are politicians, we have a platform and speak out," she said. "But meanwhile, there are millions of women victims of these criminal acts." Shultz notified each affected MP directly and provided guidance on takedown routes.
The sites themselves fell into several categories. Some hosted active deepfake video content. Others carried indexed pages where content had been removed but the target's name and photo remained. A third category — the one Shultz flags as most operationally dangerous — assembled MP biographies with official photos, social media handles, and sometimes personal information, then paired those pages with direct links to nudifier apps or the site's own image-generation tools.
“This discourages women from taking office in the first place and also crowds these women out of public life”— Benjamin Shultz, Lead researcher, American Sunlight Project
Shultz describes the flow as three clicks from an MP's profile page to a generated image. He argues that hosting the pipeline is functionally equivalent to hosting the output. Some pages went further and included sections rating the women's bodies alongside the biographical data.
Deepfake abuse has scaled fast since AI-generated synthetic media first emerged at the end of 2017. Henry Ajder, an AI adviser who has tracked the deepfake ecosystem since 2018, said the operators run these platforms as commercial businesses with the infrastructure built for volume. The same tooling has already surfaced in cases at nearly 100 schools worldwide, where students have used nudifier apps against classmates.
“These platforms are designed to run like businesses”— Henry Ajder, AI adviser and deepfake expert
Regulation has lagged the technology by years, but the last several months have moved. The UK and EU are drafting bans on nudify services. The US Take It Down Act has already been used to force at least two major deepfake sites offline, and Shultz's earlier 2024 analysis of US senators and members of Congress showed a significant drop in depictions of American politicians after those takedowns. Ahead of the European report's publication, at least half a dozen of the 160 domains in the dataset had also been removed.
Sarah Dobbe, a Dutch Socialist Party MP who was affected, said the priority is fast removal paired with support for victims outside politics — the schoolchildren and private women who make up the bulk of the target set but lack the platform to force takedowns. Kröger echoed the point, arguing that removal alone isn't enough without binding regulation on the underlying generation tools.
The counterweight to the takedown momentum is architectural: the models powering nudifier services are open-weight and freely distributed, and shuttering a hosting domain does not remove the capability from the internet. New sites replace old ones within weeks, and the search-engine indexing that Shultz relied on to map the ecosystem also gives new entrants a distribution channel. Legislation targeting the consumer-facing sites addresses the shop window, not the supply chain.
The through line for AI policy is that generative image models have produced a persistent, gendered harm at scale, and the enforcement response is finally starting to bite at the domain level even as the generation layer remains untouched. Expect the next round of EU rulemaking to push past hosting bans toward liability for the model providers and API resellers powering nudifier pipelines — that is where the economics of the abuse actually sit, and it is the fight the AI industry has so far mostly avoided.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
