Sexually Explicit Deepfake Sites Target 100-Plus Politicians in Europe

Generating nonconsensual explicit images of individuals can now be done with minimal effort, directly impacting public figures and potentially deterring women from public service.

Researcher Benjamin Shultz identified over 160 domains hosting sexually explicit deepfakes, with at least 138 female European Union Members of Parliament (MPs) appearing on these sites compared to only nine male MPs. The research indicates women MPs are 33 times more likely to be targeted. The identified sites not only host deepfake content but also provide tools and links to generate explicit images of individuals, often including their personal information and social media profiles. The study, published by Agora Digitale Transformation, analyzed nearly 5,800 MPs across 27 EU countries. Politicians from Germany, the Netherlands, Italy, and France were most frequently targeted, with seniority correlating to increased likelihood of being targeted. Shultz has individually alerted affected MPs and offered guidance on content removal, noting that the technology has become more accessible and realistic since its emergence in late 2017, leading to widespread abuse.

The scale of deepfake abuse targeting female politicians highlights a concerning trend of online intimidation and harassment, suggesting that these platforms are deliberately structured to facilitate and profit from such activities. The ease with which personal information and AI generation tools are linked on these sites demonstrates a sophisticated infrastructure designed for widespread exploitation, making it difficult to combat effectively. This systematic targeting not only causes significant psychological and reputational damage to the individuals involved but also poses a threat to democratic processes by discouraging women from participating in public life.

While some legislative efforts are underway in the UK and EU to ban related services, the slow pace of regulation means that deepfake technology continues to evolve and be weaponized. The success of takedown efforts against some sites, evidenced by a decrease in depictions of targeted US politicians, suggests that coordinated legal and enforcement actions can yield results. However, the continuous emergence of new domains and the inherent accessibility of the underlying AI technology present an ongoing challenge, requiring robust and adaptable strategies to protect individuals and uphold democratic integrity.