Between 2018 and 2025, MrDeepFakes.com accumulated over 2.2 billion views while hosting nearly 70,000 non-consensual deepfake videos. For seven years, this platform operated in near-total impunity, until an international journalistic investigation brought it down. Paris Hilton, one of its most high-profile victims, chose to transform her own experience into public advocacy. A closer look at a case that is reshaping the stakes of digital identity.
MrDeepFakes: An Industrial Platform for Non-Consensual Deepfakes
MrDeepFakes.com was not a simple dumping ground for illicit content. It was an organised infrastructure for the production and distribution of non-consensual deepfake pornography: tutorials teaching users how to generate synthetic videos using open-source tools, and a marketplace where users commissioned personalised deepfaked content targeting named individuals. Celebrities, yes but above all ordinary women: colleagues, ex-girlfriends, classmates. Some creators charged hundreds of dollars per custom video. The site also hosted more than 1,000 videos depicting violent abuse. A community of over 650,000 members gathered to share techniques, files, and misogynistic comments about their victims.
All of this over seven years. From 2018 to May 2025.
Paris Hilton has publicly stated that more than 100,000 AI-generated intimate images depicting her were created without her consent. For her, this is not an abstract discovery. She had already experienced, in the early 2000s, the non-consensual distribution of a sex tape, a first violation of her digital identity, which contributed to making her a household name before she was ready, and not in the way she would have chosen. Twenty years later, she watches the same mechanism repeat itself on an industrial scale, fuelled by generative AI.
How an International Investigation Identified the Administrator of MrDeepFakes
In May 2025, a media consortium (CBC News, Bellingcat, Danish newspaper Politiken, and investigative outlet Tjekdet) published a groundbreaking international investigation into non-consensual deepfake pornography. After three years of work, journalists traced an anonymous operator known under the pseudonym “dpfks” to his real identity: David Do, a 36-year-old Canadian pharmacist employed at the Oak Valley Health hospital network in suburban Toronto.
The identification relied on open-source intelligence (OSINT) techniques: cross-referencing massive data breaches, recurring usernames, unique passwords, IP addresses, more than a decade of digital traces reconstructed piece by piece. Before the investigation was even published, a journalist had physically shown up at the hospital. Do responded that he knew nothing about MrDeepFakes, then shut his office door.
On 4 May 2025, four days before the investigation was published, MrDeepFakes displayed a terse message on its homepage: immediate and permanent closure. Do had been notified he was about to be named. On 15 May, Oak Valley Health confirmed he was no longer employed by the network. As early as May 2025, a Danish member of parliament had called for his extradition, followed in August 2025 by Dutch MPs, arguing that numerous victims were located in their respective countries. The case is still ongoing.
“Searching for Mr. Deepfakes”: Paris Hilton and Laurie Segall Fight Back on TikTok
On 27 May 2026, Paris Hilton and journalist Laurie Segall (former CNN technology correspondent), who spent three years investigating this case, launched Searching for Mr. Deepfakes, a 14-episode documentary series released exclusively on Paris Hilton’s TikTok account. Co-produced by Mostly Human Media and 11:11 Media, the series totals approximately 40 minutes of content, with each episode running between one and five minutes.
The choice of TikTok was deliberate and strategic. Laurie Segall consciously bypassed traditional formats to reach the audience most exposed to non-consensual deepfakes: young women, who are statistically the primary targets of AI-enabled digital abuse. The episodic short-form narrative follows a “rabbit hole” logic, each episode pulls the viewer into the next. Segall described the series as a “beta test for a new journalistic playbook“: investigative depth, algorithmic format, and advocacy built in through Paris Hilton’s public profile on the DEFIANCE Act. A four-part podcast series was also launched on 4 June on Mostly Human.
For Paris Hilton, the commitment is as personal as it is political. “This isn’t about technology, it’s about power,” she declared at a Capitol Hill press conference in January 2026. “It’s about using someone’s image to humiliate them, silence them, strip them of their dignity. Victims deserve more than after-the-fact apologies. They deserve justice.”
The Legal Framework Against Intimate Deepfakes in the United States in 2026
The MrDeepFakes affair contributed to accelerating a major legislative awakening in the United States. The TAKE IT DOWN Act (Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks Act), signed into law by President Donald Trump on 19 May 2025, is the first federal US law criminalising the distribution of non-consensual intimate imagery, whether real or AI-generated. It requires online platforms to remove any non-consensual deepfaked content within 48 hours of receiving a valid notification, including identified copies. Having come into full effect on 19 May 2026, exactly one year after its signing, it now exposes platforms to fines of $53,088 per violation left unaddressed within the deadline. The FTC (Federal Trade Commission) issued warning letters to twelve major platforms on its very first day of enforcement, including Meta, TikTok, X, and Reddit.
In its wake, the DEFIANCE Act (Disrupt Explicit Forged Images and Non-Consensual Edits Act) was passed unanimously by the US Senate on 13 January 2026. It would give victims of non-consensual deepfake pornography a federal civil right to sue creators and distributors, with a minimum of $150,000 in damages per violation. As of 30 May 2026, the bill is still awaiting a vote in the House of Representatives. Paris Hilton testified in person on Capitol Hill alongside Alexandria Ocasio-Cortez and Laurel Lee to advocate for its passage, describing deepfakes as “the newest form of victimisation” in the age of generative AI.
Content Certification: The First Line of Defence
What the MrDeepFakes affair reveals is not only the violence that individuals can inflict under cover of digital anonymity. It is the structural vulnerability of any unprotected online identity. The documented victims of the site were, in their majority, ordinary women, not celebrities. People whose digital image had been captured somewhere, without their knowledge, with no means of proving the authenticity or origin of content depicting them.
In this context, preventive content certification is not a tool reserved for major brands or public figures. It is the technical response to a reality that MrDeepFakes has made impossible to ignore: in a world where any digital content can be falsified by AI, the only enforceable proof of authenticity is one that has been established at the source, before the attack.
The law steps in after. Certification steps in before.
