Hilton series targets deepfake abuse

Paris Hilton and journalist Laurie Segall have launched a 14-part TikTok investigation that names a Canadian pharmacist linked to MrDeepFakes, the shuttered website that became one of the world’s largest hubs for non-consensual AI-generated explicit videos.

The series, Searching for Mr. Deepfakes, began streaming on Hilton’s TikTok channel on 27 May and follows Segall’s multi-year effort to identify the people behind the site. It places fresh public attention on David Do, a former pharmacist in the Greater Toronto Area who was identified last year by a cross-border journalism investigation as a key figure in the administration of MrDeepFakes.

MrDeepFakes operated from 2018 until May 2025, when it went offline after a message on the site cited the loss of a critical service provider. During its seven-year run, the platform is estimated to have accumulated about 2.2 billion views and hosted tens of thousands of non-consensual AI-generated explicit videos. Its forums allowed users to request, trade and commission synthetic sexual content, including material targeting celebrities and private individuals.

Hilton’s involvement gives the project a high-profile advocate at a time when lawmakers, regulators and technology platforms are under pressure to confront AI-enabled sexual exploitation. She has said that more than 100,000 explicit deepfake images of her have been created without her consent, placing her among the public figures most visibly affected by the spread of synthetic intimate imagery.

The TikTok format marks a shift in how investigative work on technology abuse is being presented to mass audiences. Rather than releasing a traditional documentary or long-form report, the project uses short vertical episodes designed for social media distribution. The approach reflects a wider effort by journalists and advocacy groups to reach younger users on the same platforms where manipulated images and videos can spread quickly.

Do has been described as a former hospital pharmacist connected to Oak Valley Health in Ontario. After his name was linked to the site last year, he was placed on leave and then left the organisation. The Ontario College of Pharmacists opened a separate review. The reporting that identified him relied on open-source research, leaked material, online handles, domain records and platform activity. He has not publicly responded to the allegations in detail.

The case has exposed the uneven legal landscape surrounding deepfake abuse. Some jurisdictions have laws covering non-consensual intimate images, but enforcement becomes more difficult when the material is synthetic, hosted across borders or circulated through anonymous accounts. Victims often face a heavy burden in proving harm, identifying perpetrators and securing removal from platforms that may be outside their home country.

The United States has moved further into federal enforcement with the TAKE IT DOWN Act, signed into law in May 2025. The law criminalises the publication of non-consensual intimate imagery, including AI-generated material, and requires covered platforms to remove flagged content within defined timelines. Federal Trade Commission enforcement began this month, giving regulators a new mechanism to act against services that fail to comply.

Another proposal, the DEFIANCE Act, seeks to give victims the right to sue creators and distributors of non-consensual sexually explicit deepfakes. It has bipartisan support and has passed the Senate, but its final path depends on action in the House. The measure would add a civil remedy to the criminal and platform-removal framework already in place, addressing criticism that victims need direct legal tools against individuals who create and circulate such material.

Researchers tracking the deepfake ecosystem warn that shutting down a major site does not necessarily eliminate the abuse. Studies of online forums indicate that activity can migrate to other platforms after enforcement shocks, with users shifting across websites, private groups and model-sharing communities. The availability of open-source face-swapping tools and downloadable AI models has made it easier for non-specialists to generate realistic images using limited source material.

The problem is no longer confined to celebrities. Schools, workplaces and local communities have reported cases involving classmates, colleagues and private citizens. The harm can include reputational damage, harassment, emotional trauma and professional consequences, even when viewers understand that the images are fabricated.

Technology companies face scrutiny over detection, reporting channels and search visibility. Some platforms have strengthened takedown systems and restricted search results for explicit deepfake material, but advocates say victims still encounter slow response times and repeated reuploads. Detection tools can help, but they remain imperfect as generative systems improve and manipulated content is resized, altered or shared in closed communities.

Hilton and Segall’s series arrives as the public debate shifts from whether deepfake abuse should be treated as a serious offence to how governments and platforms can enforce accountability across borders. By naming an alleged operator and tracing the infrastructure behind MrDeepFakes, the project seeks to turn a largely anonymous online economy into a subject of public scrutiny.