New research has found 50 Members of the European Parliament caught in the online ecosystem of sexualised AI deepfakes, including 46 women. As 119 MEPs demand tougher enforcement, researchers warn that the technology is becoming cheaper, more decentralised and increasingly profitable.
One in six women serving as members of the European Parliament was found to be featured on websites facilitating sexualised deepfakes, according to new research by think tank Agora Digitale Transformation.
The organisation examined the names of all 720 MEPs against 200 websites known to host or facilitate AI non-consensual images. They identified 50 MEPs, 46 women and four men. Women lawmakers were estimated to be 33 times more likely than their male counterparts to be affected.
The phenomenon extends far beyond politics. The number of AI non-consensual videos increased from roughly 500,000 in 2023 to eight million in 2025. This means a 1,500% rise in just two years. Researchers identified 88 publicly accessible sites. Almost all, 93%, could be reached through Google search.
Just imagine for one moment going in the morning to work, and then opening your laptop, and you learn there are actually naked pictures of you published on the internet. But surprisingly, you never took them.
— MEP Lena Schilling (Greens/AT)
For Austrian Green lawmaker Lena Schilling, the issue became personal. “Just imagine for one moment going in the morning to work, and then opening your laptop, and you learn there are actually naked pictures of you published on the internet. But surprisingly, you never took them. That’s what happened to many, many other women and to me. And it’s a horrible thing,” she described.
Schilling warned that the phenomenon could discourage women from entering public life. “If we are facing a situation where women who run for office are also thinking: ‘If I run for office, maybe there will be naked pictures of me on the internet.’ This is intimidating”, she worries.
119 lawmakers demand action
A group of 119 lawmakers from the EPP, Socialists and Democrats, Renew Europe, The Left and Greens signed a letter calling on the European Commission and EU governments to ensure Europe’s new prohibition on AI nudification tools is enforceable. The rule, which is a part of the AI Act, becomes applicable on 2 December.
The lawmakers are also calling for stronger enforcement of the Digital Services Act against platforms and search engines that distribute or index non-consensual material. They want the Commission to act more quickly in its investigation into X and Grok. They also demand earlier national implementation of the EU directive on combating violence against women.
Reacting to the letter, a Commission spokesperson said such material was illegal and had no place in Europe, adding that Brussels would be ready to enforce the new rules “from day one”.
The debate, however, highlights a broader gap between legislation and enforcement. Particularly when responsibility for removing illegal material is spread across platforms, search engines and national authorities.
In comparison, the United States’ TAKE IT DOWN Act requires covered platforms to remove non-consensual intimate material, including synthetic ones, within 48 hours of receiving a valid request.
Researcher Benjamin Shultz said the two-day deadline offered Europe a model worth considering. But neither approach eliminates the burden on victims. Under the US system, the process still begins when a victim discovers the content and files a complaint.
Shultz argued that European regulators should increasingly act on their own initiative rather than relying on victims to identify and report illegal material.
Follow the money
The problem is also becoming harder to address as the technology becomes more decentralised. Shultz said researchers are increasingly seeing decentralised online forums where users train AI models on one specific person’s face, share those models with others, and generate images locally on their own devices.
In such cases, conventional platform enforcement becomes extremely difficult.
His proposed response is to move “upstream”, towards the financial and technical infrastructure supporting the industry. Nudifier services need payment processors, advertising, app stores, domain registrars, hosting providers and search engines.
The Agora think tank argues those intermediaries should face greater responsibility for facilitating or profiting from AI sexual images. Its study says such services currently form part of a lucrative ecosystem while facing limited responsibility for their role in distribution.
What regulators still do not know
Important questions also remain about how sexualised deepfakes are being used.
One is whether they are deliberately deployed as part of coordinated campaigns to discredit public figures. Researchers have identified forums where users requested material involving outspoken politicians. Schilling said one image involving her appeared to have been generated from footage of a press conference on climate issues.
The Internet Watch Foundation reported that it identified 6,310 AI-generated images in the first six months of 2026. In the whole of 2025, it was 4,512. Girls appeared in 98% of the images, while the organisation also documented an increasing number involving younger children.
The 2 December deadline will provide an early test of whether the EU’s new rules amount to more than simply declaring the practice illegal. And whether victims will actually see a difference in how quickly the material is identified and removed.
