While how each app or website works might vary, all of them offer this similar service. Although the manipulated image isn’t actually showing the victim’s real nude body, it can imply this.

Perpetrators who use undress AI tools might keep the images for themselves or might share them more widely. They could use this images for sexual coercion (sextortion), bullying/abuse or as a form of revenge porn.

Children and young people face additional harm if someone ‘undresses’ them using this technology. A report from the Internet Watch Foundation found over 11,000 potentially criminal AI-generated images of children on one dark web forum dedicated to child sexual abuse material (CSAM). They assessed around 3,000 images as criminal.

The IWF said that it also found “many examples of AI-generated images featuring known victims and famous children.” Generative AI can only create convincing images if it learns from accurate source material. Essentially, AI tools that generate CSAM would need to learn from real images featuring child abuse.

One undress AI site says that their technology was “not intended for use with male subjects.” This is because they trained the tool using female imagery, which is true for most of these types of AI tools. With the AI-generated CSAM that the Internet Watch Foundation investigated, 99.6% of them also featured female children.

Research from Graphika highlighted a 2000% increase of referral link spam for undress AI services in 2023. The report also found that 34 of these providers received over 24 million unique visitors to their websites in one month. They predict “further instances of online harm,” including sextortion and CSAM.

Perpetrators will likely continue to target girls and women over boys and men, especially if these tools mainly learn from female images.

However, the Ministry of Justice announced a new law this week that will change this. Under the new law, those creating sexually explicit deepfake images of adults without their consent will face prosecution. Those convicted will also face an “unlimited fine.”

This contradicts a statement made in early 2024. It said that creating a deepfake intimate image was “not sufficiently harmful or culpable that it should constitute a criminal offence.”

As recently as last year, perpetrators could create and share these images (of adults) without breaking the law. However, the Online Safety Act made it illegal to share AI-generated intimate images without consent in January 2024.

Generally, this law should cover any image that’s sexual in nature. This includes those that feature nude or partially nude subjects.

One thing to note, however, is this law relies on the intention to cause harm. So, a person who creates a sexually explicit deepfake must do so to humiliate or otherwise harm the victim. The problem with this is that it is fairly difficult to prove intent. As such, it might be difficult to actually prosecute parties creating sexually explicit deepfakes.

For students under the age of 18, most schools must notify parents and carers. They must also report the case to the local safeguarding body. In some cases, school police liaison officers will also advise the school.

Training for relevant school staff will vary by region. Still, most professional training prioritises themes such as:

• the legal implications of child-generated sexual images;
• student wellbeing; and
• assessment of the incident for factors related to safeguarding such as the relationship between the involved parties, coercion and age-disparity.

To do so, however, is to misunderstand the power dynamic that often exists behind these activities.

Sexting sits alongside a number of other behaviours such as sextortion and ‘revenge porn’ that exist in cultural context. These things often normalise and glamorise sexual exploitation.

Messages about the legal and safety implications of child-generated sexual images are extremely important. However, teachers and other professionals who work with children and young people should include discourse about not requesting nudes.

Furthermore, they should create opportunities to explore the underpinning themes of exploitation and pornography, healthy relationships, and respect and care for others.

Read the full report

Other supporting resources

Y99 is an online chat room service from India that lets people anonymously chat with others. It mimics the chat rooms of the 1990s and early 2000s. Users only need a username to talk with others. Users can then enter chat rooms to talk with strangers via text or voice, or they can share photos and links to videos.

Some chat rooms require account verification instead of just a username. This means users must sign up and verify their account to join a chatroom.

In addition to chat rooms, users can privately chat with strangers. They can do this by choosing a user and starting a chat, or by using the ‘random’ feature on Y99. Like Omegle, the site then randomly matches you to another user.

Users can also create their own public or private chat rooms. They can choose to add a password or untick the option to ‘submit to Y99’s Search’.

Like chat rooms of the ’90s and ’00s, the website is chaotic and filled with ads, users and potential bots.

Why users might like Y99

Y99 has the nostalgic factor for some older users who chat on it. However, teens might like the ‘retro’ style, variety of chat rooms and the anonymity they can’t get from regular social media.

Chat rooms cover a variety of themes and topics — from ‘Teens Only’ to LGBT rooms and support for depression or other issues. As such, teens might find a community they identify well with. However, other options like Tumblr or Discord can offer similar, safer experiences.

Y99’s safety features are minimal. It includes reporting and blocking features but little else.

The site doesn’t have clear age requirements to use the site. However, its privacy policy states that it does not ‘knowingly’ collect personal information from children under 13.

This does not mean the site is appropriate for children over the age of 13. In fact, it might mean the opposite. Y99 might collect personally identifiable information from anyone aged 13 or older. Additionally, it’s easy for users to click on the wrong thing and end up following an ad.

While Y99 does feature some age-restricted rooms for ‘teens only’, a user need only lie about their age to gain access. It’s impossible to know for sure that users in the chatroom are teenagers.

For the above reasons and other potential issues, we recommend those under the age of 18 to avoid using Y99. Instead, they should use more well-established communication platforms and social network sites.

See age-appropriate social networking platforms.

If your child uses or wants to use Y99, it’s important to speak with them about it. Additionally, if you do let your child use the platform, you should moderate the use closely and regularly check in.

When discussing Y99 with your child, here are some questions to ask.

• Why do you want to use Y99? Understanding where their interest comes from can help you offer alternatives. Tumblr, Discord and even Reddit could offer alternatives to finding a community. If they’re more interested in the ‘instant’ aspect, they might be able to find an active Discord server to do the same.
• What are the risks of using Y99? Let young people explain the risks and acknowledge potential harm. If they’re unable to recognise these risks, they might not be prepared to use the site. This kind of conversation can also help you identify missing knowledge.
• What would you do if…? Talk about potentially harmful scenarios like receiving an inappropriate image or feeling uncomfortable because of a stranger. Explore the actions they know to take. Again, if they’re not sure, this might not be the right site for them.
• Is it okay if…? Think about scenarios such as if a stranger asks for personal information or to meet on a different platform. Talking about what is and isn’t okay can help young people spot unhealthy behaviour immediately.