Online communities operating across platforms such as 4chan, Reddit and private messaging applications are helping users create and distribute AI-generated nude images without consent, according to a new study by researchers at the University of Florida.

To advertise here,

The research has uncovered an extensive online ecosystem in which individuals exchange AI tools, technical knowledge and resources to produce synthetic non-consensual intimate imagery (SNCII), raising serious concerns about digital abuse, privacy and the exploitation of women and other vulnerable individuals.

Led by Kevin R. B. Butler, professor and chair of the University of Florida's Department of Computer & Information Science & Engineering (CISE), the study examines the networks sustaining so-called nudification applications, which use artificial intelligence to digitally remove clothing from photographs of people.

The findings come as Domestic Violence Awareness Month concludes, drawing attention to the growing use of AI-generated intimate images as instruments of harassment, intimidation and coercive control.

“There's a strong link between nonconsensual intimate imagery and domestic violence,” Butler said.

He warned that such content could be weaponised against victims, particularly in abusive relationships.

“Real or synthetic imagery can be used to threaten exposure, known as ‘sextortion,'” Butler said. “This type of imagery can be used to prevent a partner from leaving, reporting violence or seeking custody of children because of the perceived reputational harm.”

Researchers analyse over 3.6 lakh online posts

The study, titled It Takes a Forum: Resources & Stakeholders Sustaining the Online Synthetic Non-Consensual Intimate Image Creation Ecosystem, investigates how online communities facilitate the creation and circulation of AI-generated intimate images.

The research team examined 282,154 posts on 4chan and 78,308 submissions on Reddit published between June 9 and November 21, 2025.

Researchers initially conducted detailed manual analyses of 200 randomly selected posts from each platform to identify patterns involving participants, activities and resources.

They then developed a behavioural classification framework, which was applied to the larger datasets to understand how users acquire knowledge, exchange tools, and participate in the creation of non-consensual imagery.

The findings revealed differences in how individuals use various online platforms.

Technically sophisticated participants were more likely to engage with communities such as 4chan, while less experienced users were more active on platforms including Reddit.

However, the researchers found that the ecosystem extended beyond public discussion forums, with participants relying on private communication channels to exchange resources and evade restrictions.

'Educators' help newcomers create AI-generated nude images

One of the study's most significant findings was the identification of individuals described as "educators", who help inexperienced users acquire the knowledge needed to create synthetic intimate images.

These individuals provide technical guidance, troubleshooting assistance, tutorials and access to resources, effectively helping newcomers become active participants in the ecosystem.

According to the researchers, 56.3 per cent of the identified educators helped less experienced users progress towards more advanced levels of participation.

The study also uncovered what researchers described as a shadow infrastructure that adapts to enforcement measures.

Private messaging platforms, including Telegram and Session, were used to exchange tutorials, AI models and other resources associated with the creation of non-consensual imagery.

Researchers found that this interconnected system allowed participants to continue sharing information even when restrictions were introduced on individual platforms

Users appear undeterred by legal consequences

The study also identified a concerning gap between awareness of legal restrictions and the behaviour of individuals involved in creating and sharing AI-generated intimate images.

Researchers observed that several participants appeared familiar with regulations prohibiting such activities but showed little concern about potential consequences.

“The fact that they were so open about their sharing of resources and images was one of the things that motivated us to begin this study,” said Bernardo B. P. Medeiros, a CISE doctoral student and co-lead author of the paper.

“We began this work soon after the TAKE IT DOWN act, but despite this, we found that the 4chan community was not deterred from posting images. Even today, people are still posting these kinds of images on 4chan.”

The findings come against the backdrop of legislative efforts in the United States to address the growing threat of non-consensual intimate imagery.

The TAKE IT DOWN Act, enacted in 2025, introduced federal measures targeting the publication of non-consensual intimate images, including AI-generated content, and established procedures for their removal from online platforms.

However, the researchers' observations suggest that legal measures alone may not be sufficient to dismantle the networks facilitating the creation and circulation of such material.

Study identifies ways to disrupt AI nudification networks

Rather than focusing exclusively on individuals who generate or distribute abusive imagery, the researchers examined the broader infrastructure enabling these activities.

The study identified several potential intervention points that could help disrupt the ecosystem.

One approach involves identifying and targeting the influence of individuals who provide technical training and resources to newcomers.

By limiting the ability of these educators to recruit and guide new participants, researchers believe it may be possible to reduce the spread of knowledge used to produce non-consensual imagery.

Another proposed intervention focuses on major public repositories that host AI image-generation models and related resources.

The researchers suggested that companies operating these platforms could introduce stronger restrictions on materials associated with non-consensual image generation.

Such policies could provide measurable opportunities to assess whether platform-level enforcement reduces the availability and circulation of harmful resources.

Researchers avoided viewing or storing abusive images

The research team emphasised that strict ethical safeguards were followed throughout the investigation.

Although researchers collected publicly available discussion data from Reddit and 4chan, they did not interact with participants or contribute to communities associated with illicit activities.

All collected data was anonymised, and the researchers neither viewed nor stored the intimate images discussed in the forums.

The study was conducted by Butler alongside University of Florida doctoral researchers Bernardo B. P. Medeiros, Malvika Jadhav and Allison Lu, CISE faculty member Vincent Bindschaedler, and Georgetown University researcher Tadayoshi Kohno.

Their paper is scheduled to be presented in November at the Association for Computing Machinery's 2026 Conference on Computer and Communication Security in The Hague, Netherlands. A preprint has been made available ahead of the conference.

The researchers hope their findings will help policymakers, technology companies and online platforms identify weaknesses in existing safeguards and develop more effective measures to prevent the creation and circulation of non-consensual intimate imagery.

The study also highlights the need to address not only the harmful content itself but the online communities, technical resources and support networks that enable its production.