Deepfake Nudes & Young People: Navigating a New Frontier in Technology-facilitated Nonconsensual Sexual Abuse and Exploitation

Rapid & scalable deepfake nudes

While photo manipulation tactics have long been used to create or alter child sexual abuse imagery, earlier methods were typically crude — requiring significant time, skill, and technical expertise to produce photorealistic outcomes.

Generative AI has fundamentally changed this dynamic, enabling the rapid and scalable creation of highly realistic content with minimal skill or effort.

Unlike traditional forms of photo manipulation software, deepfake technologies uniquely empower the creation of explicit content targeting anyone — regardless of consent, prior digital behaviors (e.g., sharing intimate imagery), or even knowledge of the content’s existence.

Creators

Adults with sexual interests in children may use AI to fabricate abusive images of children both known to them and not, while others exploit deepfakes for sextortion, threatening to release fabricated images unless the victim complies with their demands.

Some young people themselves may also create deepfake nudes of their peers out of sexual curiosity, peer pressure, or as a form of bullying.

Risks & challenges

For those working in child safety, deepfake technologies introduce significant new risks and challenges in the fight against technology-facilitated child sexual abuse, many of which  and undermine the ecosystem’s ability to respond and prevent abuse.

Some of these new challenges represent technical hurdles and are inextricably tied to the technologies’ ease of access, lack of unified implementation of available safeguards (e.g., safety-by-design principles), and capacity to produce highly photorealistic content at scale.

Other new challenges reflect systemic issues. While progress has been made to advance trauma-informed responses to victim disclosures of sexual abuse and exploitation, significant barriers persist, including shame, humiliation, fear of not being believed, and self-blame.

The growing prevalence of deepfake nudes, especially in youth spaces like schools and social media, underscores the urgency for immediate and systemic intervention alongside technical ones.

Exploring adult apps

Despite age restrictions, a substantial proportion of teens (aged 13-17) reported accessing platforms exclusively designed for adult users (aged 18 or older).

Approximately 1 in 5 (22%) teens reported using a dating app, 1 in 10 (10%) had accessed OnlyFans, and nearly 1 in 4 (23%) had visited pornography sites.

Sharing sexual imagery

Young people were asked about their experiences sharing sexual imagery of themselves while under 18, with 1 in 6 (18%) reporting they had. Among teens, 1 in 8 (13%) reported having shared sexual imagery of themselves with others.

Past research has established that while many report sharing the imagery with someone they know offline, they also report sharing the content with other users who they only know online. That research also established that minors who share their nude imagery report doing so with other people across a spectrum of ages, including other minors, adults, and people whose ages they don’t know.

Online solicitations

Roughly 1 in 3 young people (36%), reported being asked to share sexual imagery of themselves while they were under the age of 18 by someone they had met online.

LGBTQ+ respondents (53%) and women and girls (41%) were the most likely to report having this experience. Rates of the experience appeared to increase with age.

Awareness

Overall, 41% of young people indicated they had heard of the term deepfake nudes, including nearly 1 in 3 (31%) of teens. Men and boys were more likely to have heard the term than women and girls (+9%). LGBTQ+ young people were also more likely to have heard the term than their non-LGBTQ+ peers (+20%).

Young people reported that their peers were encountering a mixture of experiences with deepfake nudes as minors– both being targeted with abusive images and as people involved in creating or redistributing those images.

Around 1 in 8 young people indicated that they personally knew someone who, while under 18, had been the target of deepfake nudes (13%) and/or someone who had used the technologies to create or redistribute deepfake nudes of minors (12%).

In total, 1 in 6 (17%) respondents indicated they personally knew someone who had encountered at least one of the following experiences: 

1. Had a deepfake nude made of them by someone else
2. Discovered a deepfake nude of them was being (re)shared
3. Created a deepfake nude of themselves
4. Created a deepfake nude of a minor
5. Reshared a deepfake nude of a minor

While the comparative base size was small, LGBTQ+ teens were much more likely than their peers to indicate that they knew someone impacted; 18% reported knowing someone targeted by the technologies, and 23% reported knowing someone who used the technologies to create or distribute the content of others.

Perception of harm

Overwhelmingly, young people (84%) indicated they believed the content harmed the person depicted in deepfake nudes.

Among those who believed deepfake nudes were harmful, the top three reasons they identified were:

1. The emotional and psychological impacts experienced by the victim(s) (31%)
2. The potential for reputational damage of the victim(s) (30%)
3. The inability for viewers of the deepfakes to recognize the content is not authentic (25%)

By comparison, around 1 in 6 (16%) young people believed deepfake nudes either were not harmful to the person depicted or that it depended on the situation.

Among young people who did not think deepfake nudes caused harm to the victims, the top two leading reasons were because they saw the imagery as “fake” and/or “not real” (28%) or because no physical harm was involved (7%).

When the portion of young people who indicated they thought deepfake nudes were either harmless or situationally dependent were asked what would make them more likely to recognize harm, the most influential factor was if viewers believed the deepfakes were authentic (40%).

One-third said they would be more likely to think deepfake nudes were harmful if the images couldn’t be removed from the internet. Similarly, 29% said sharing or distributing the content would make it more harmful — whether shared with people who knew the person in the image or posted on digital platforms.

Perception of legality

Young people were presented with three different scenarios involving a minor creating deepfake nudes: of themselves, of someone else under the age of 18, and of someone else aged 18 or older.

In total, most young people recognized the behavior as illegal across all three scenarios.

Young people were slightly more likely to believe that a minor creating an image of themselves was legal (23%) than of someone else (19%), regardless of the other person’s age. For some, they thought the legality was dependent on the situation involved.

Responding to victimization

Among the portion of young people who had experienced deepfake nude victimization (regardless of whether it happened while they were a minor or young adult at the time of victimization), the majority (84%) sought some form of support, using either online safety tools or seeking offline forms of support.

Around 1 in 6 (16%) deepfake nude victims indicated they did not seek support at all — either because they ignored it or were unsure in their response. This rate is similar to rates observed in previous Thorn research looking at minors’ responses to potentially harmful online sexual interactions.

Among young people who’ve been victimized by deepfake nudes, over half (60%) used at least one online tool, such as platform-based blocking or reporting features, as part of their response, and more than half (57%) sought at least one form of offline support, by telling someone about what had happened to them, whether it be a parent, school authority, friends, or police. While the sample size was small, nearly half (48%) of teens who had deepfake nudes created of them indicated they told their parents/guardians or a trusted family member.

Blocking technologies were the most commonly utilized online tool, with nearly half (48%) of victims choosing to block the offending user, compared with 35% who reported the user. 

In fact, blocking the other user emerged as the most frequently reported response type, followed by reporting the user and telling a parent or other trusted family member (34%). Among teens, blocking the other user (49%) and telling parents (48%) were the most reported response types.

Response versus anticipated response

As with other online harms there was a notable gap between the anticipated responses of nonvictims of deepfake nudes and the actual responses of those who had been victimized.

Among those who have not personally experienced deepfake nude abuse they were most likely to think they would tell a parent or trusted adult (62%), while only 1 in 3 (34%) victims of deepfake nude abuse did. This gap was also seen among teens: 72% anticipated they would tell a parent or trusted adult, while only 48% of teen victims did. 

Other discrepancies between anticipated response and actual response were also noted, including a decreased likelihood for victims to report the offender to the police (-32%) or to the platform where the content is being shared (-22%). 

Alternatively, victims were more inclined to ignore the experience (+12%) or confide in online friends (+8%) when compared to nonvictims’ anticipated responses.

Deepfake nude creators

Among young people surveyed, 2% indicated they had used technology to create deepfake nudes of another person.

While the sample size was small (n = 24), it offers an initial snapshot into the practices and motivations of young people involved in creating deepfakes.

The data on young deepfake creators explored in this section should be viewed as directional, underscoring the urgent need for more nimble and comprehensive research on young people’s role in perpetrating this abuse.

Victims targeted

Among the subsample, deepfake creators reported being most likely to have created deepfake nude imagery of an adult (62%), with roughly 1 in 3 (36%) indicating they had made deepfake nude imagery of a minor.

Creators of deepfake nudes also overwhelmingly said they created the content of females (74%).

Motivations

When asked about their motivations, the portion of young people who have created deepfake nudes of others described varied and individualized reasons, including sexual curiosity, pleasure-seeking, revenge, or creating the content as a result of pressure or influence from friends.

Technologies used

When asked how they learned about the websites, platforms, or apps they used to create the content, deepfake nude creators described multiple, highly accessible pathways. This included via social media platforms (71%), such as Facebook, Instagram, Snapchat, TikTok, and YouTube, via search engines (53%), such as Google and Bing, and via direct link (25%) that someone shared with them.

Deepfake nude creators were also asked how they accessed the technologies they used to create deepfake nudes.

Most (70%) creators said they downloaded the app they used from their device’s app store (e.g., Apple’s App Store or Google’s Play Store).

Roughly one in three creators (30%) indicated they didn’t have to download the technology they used to create the content.

When asked to identify the websites, platforms, or apps they used to create the deepfake nude content, deepfake nude creators described several technologies involved in some way in the creation of the deepfake nude content. 

Sometimes, they identified specific apps publicly advertised as creating generative AI content (e.g., DeepFaceLab, AnimeGenius).

In other cases, respondents identified general-purpose technologies for image editing (e.g., Photoshop, Canva) or image sharing more broadly (e.g., Snapchat, OnlyFans, PornHub).

The responses from deepfake nude creators about the technologies they used highlight how creating deepfake nudes often involves more than just the actual generative AI technology or model — it involves sourcing benign images to seed the deepfake and leveraging platforms to distribute the resulting content.

Sharing the deepfakes they created

Deepfake nude creators were also asked if they had shared the deepfake nude content they made of someone else. 

Considering a portion of creators indicated they never shared the content with others, potentially implying that it was created solely for their personal consumption, reinforces a reality that reported victimization rates likely underestimate the prevalence of this experience. That is, not only do victims face challenges to disclosure, but also many victims may not know that the content exists.