For the first time, researchers have identified and named the key players facilitating the growing problem of AI-generated explicit sexual imagery online.
The creation and sharing of AI-generated non-consensual intimate imagery (AI-NCII) has increased rapidly due to the ease with which the content can be generated and shared via websites, social media and private messaging platforms, many with end-to-end encryption.
In 2023, estimates suggested 500, 000 such videos existed online. By 2025, this number was estimated to be 8 million, mostly targeting women and girls.
The study entitled ‘The Backbone of Abuse’, published in the Journal of Online Trust and Safety, is led by Sarah Morgan from Lancaster University with co-authors Dr Sophie Nightingale, also from Lancaster, and Professor Hany Farid of Dartmouth College.
Sarah Morgan said: “Anyone with a single photo online can be quickly and easily targeted in content that falsely depicts them in sexual or intimate contexts, of any nature, without their consent. Emerging evidence demonstrates the wide-ranging and persistent harms of this digital abuse.”
In 2024, an investigation by Channel 4 News found that almost four thousand famous people were listed on five of the most visited deepfake pornography websites,
These included female actors, journalists, musicians, TV stars and YouTubers whose faces were superimposed onto pornographic material using AI. In addition, industry analysis showed there were over four billion views of 40 of the most visited deepfake pornography sites.
This research is the first attempt to identify and name the key players facilitating —either knowingly or unknowingly— the technological infrastructure holding up sites hosting AI-NCII. Infrastructure refers to the technologies and architecture - such as website hosting, domain name servers, and content delivery networks - that enable a website to deliver content to users and without which sites cannot successfully function.
In addition, the team captured contextual data relating to each site, finding that some sites hosted hundreds of thousands of such videos, and sites primarily targeted Hollywood and Hindi-film industry stars, K-pop idols, and well-known feminists and activists, with nearly all content featuring females.
The team identified 88 sites hosting AI-NCII of which 38 exclusively hosted the content. Five key players emerged as the dominant infrastructure providers enabling the sites to function, with Cloudflare being the most recurring name.
They also assessed if any sites were found by Google Search and revealed that 93% were easily discoverable in a search.
Laws are being implemented in various countries, including England and Wales, which criminalise sharing, creating, and/or soliciting AI-NCII but lead researcher Sarah Morgan said this is not enough.
“Such legislation will offer some deterrence. But what further steps can be taken to prevent this abuse from occurring in the first place? Can we cut the distribution supply? This abuse needs to be prevented—putting pressure on the technology platforms that distribute such content to the public may be our most optimal chance of reducing the wildfire spread of AI-generated sexual abuse.
“We call on the infrastructure providers facilitating AI-NCII-hosting sites to take a firm stance on this abusive content: they should block sites that exclusively host AI-NCII and suspend services of sites that host AI-NCII within a larger offering until good-faith efforts are made to remove the abuse.
“They should deploy preventative safeguards, such as utilising perceptual hashing to extract digital signatures of confirmed AI-NCII content and stop its redistribution, and block or suspend their services for sites which are known to be hosting AI-NCII.”
Dr Sophie Nightingale adds that it is also time to stop blaming the targets of AI-NCII.
“All too often the victims of abuse that is sexual in nature are attributed responsibility—there is an urgent need to change this. Blame should be assigned to the perpetrators who create, distribute, and facilitate the sharing of fake content that has potential to ruin lives.”
Lancaster University researchers are also examining the impacts and harms of NCII. If you want to have your say on how this abuse could be tackled, or to share your own experiences, please take this confidential survey.
https://lancasteruni.eu.qualtrics.com/jfe/form/SV_0iXYjMVd1y2ctJc
Journal of Online Trust and Safety
Data/statistical analysis
Not applicable
The Backbone of Abuse’
30-Sep-2026