According to the NGO Agora Digitale Transformation, 7.3% of women elected to a European National Assembly are targeted by pornographic deepfakes, compared to just 0.2% of their male counterparts. The think tank found that European female parliamentarians face 33 times more risk of having their images manipulated or associated with sexist or sexual deepfakes than their male colleagues. This conclusion was reached after cataloging the number of elected officials in European parliaments whose images had been used to generate pornographic deepfakes or associated, in some way, with this type of AI-generated imagery. Out of 5,872 members of European parliaments, the NGO identified 138 female parliamentarians as victims of this type of image manipulation. These women are from 22 of the 27 EU member states. Among their male colleagues, who are more than twice as numerous, only nine are targeted by this type of manipulation. Agora Digitale Transformation tested the names of 94.4% of 6,218 members of the lower houses (the National Assembly in France) or national parliaments, when they are composed of a single chamber, in the 27 EU member states. After excluding vacant seats, non-binary individuals, and members not affiliated with a party, it tested the names on 160 domain names known for hosting pornographic deepfakes, "celebrity leaks," and other sexist or sexual images of women in the public space. Using queries to the Google API, it classified the positive results into six categories: whether non-consensual intimate images (NCII) were actively hosted by the website, whether an image generation tool was linked to the biographical information of a person whose name was explicitly mentioned, whether the result included raw images, whether there were traces of previous searches for the same type of content related to the person studied. The main finding was that the names of 7.3% of the 1,888 female MPs tested returned positive results, compared to 0.2% of the 3,837 male MPs. This means that one out of every 14 elected woman risks having her image misused for sexist or sexual purposes, compared to one out of every 500 elected men. Agora Digitale Transformation notes that women in positions of responsibility — party leaders, members of ministerial cabinets — are overrepresented in these image manipulations. "Achieving greater media visibility has a cost for women that men reaching similar statuses do not have to pay," the think tank states. Members of right-wing and left-wing parties are also slightly overexposed compared to their counterparts in centrist parties. The NGO also notes the existence of websites it describes as "portals" that offer official content representing a parliamentarian and link it to applications dedicated to generating sexualized or nude imagery. In its sample, it indicates that 61 websites act in this manner, directly providing the initial images that could be misused, while 36 suggest "nudefying" various parliamentarians without providing the initial images. Each time, however, they provide biographical elements to easily recognize the person and their social media accounts. These portals include AI-generated sexist or sexual representations for three parliamentarians. They sometimes provide "notations" of the bodies of the women targeted. During the investigation, the site CFake was suspended by a joint action of French and US law enforcement. In July 2026, its administrator, a 47-year-old Nice-based computer scientist, was sentenced to two years in prison with a suspended sentence, a 10,000 euro fine, the confiscation of part of his cryptoassets, and the payment of 83,000 euros in compensation to the civil parties. Agora Digitale Transformation was able, in certain configurations, to recover several dozen search indexes linking the names of parliamentarians to content removed from CFake. These always appeared in the Google Custom Search API more than a month after their removal by the authorities. While this does not necessarily mean that these results can be simply reproduced via the search engine's user interface, this finding echoes a recent article by academics Sarah Morgan, Sophie J. Nightingale, and Hany Farid, in which they point out the malfunctions in the dereferencing of illegal content by actors such as Google, Cloudflare, or Namecheap. More than calling for traditional appeals to implement safeguards in AI systems to prevent them from generating non-consensual content and to extend the use of the C2PA standard, which allows content to be traced, Agora Digitale Transformation urges working on the persistence of search indexes. As long as these indexes continue to link real people's names to sexist or sexual content, even when it has become inaccessible, they contribute to perpetuating the reputational risks and potential psychological effects suffered by the victims. The NGO also calls for implementing existing texts rather than just multiplying legislation. In the Netherlands, for example, there are no specific laws targeting deepfakes, but they are actively prosecuted using more general texts relating to image manipulation or privacy protection. The issue is all the more important as in various countries, including France, upcoming elections risk multiplying cases of manipulation of the images of political personalities.