Undress AI: What Parents and Caregivers Need to Know to Keep Children Safe

Undress AI describes a category of tools that use artificial intelligence to remove clothing from people in photos or videos. Although each app or website works differently, they all offer the same basic service. The manipulated image does not show the victim's real body, but it can still imply nudity, and perpetrators may keep these images or share them widely. They can be used for sexual coercion (sextortion), intimidation, abuse, or revenge pornography. Because the technology is marketed with suggestive language, young people may be drawn in by curiosity.

Children and young people may not yet understand the law or the difference between harmful tools and apparently harmless fun. Suggestive marketing makes these services appealing, and children may follow their curiosity without realising the consequences. A child who uses such a tool, or shares an image created with it, may be breaking the law without knowing it. Being 'undressed' by others using this technology also exposes children to serious and lasting harm.

The novelty of AI undressing tools can expose children to explicit material and encourage them to experiment. Because the result is not a 'real' nude image, children may wrongly believe that making or sharing these images is acceptable. If they share the image with friends 'as a joke', they may be committing an offence. Without the guidance of a parent or caregiver, they may continue this behaviour even when it harms others.

Many legitimate generative AI tools require payment or subscription. Free 'deepnude' websites may produce poor-quality results or have weak security. If a child uploads a photo of themselves or a friend in ordinary clothing, the site or app could misuse it, including the deepfake nude it creates. Children are unlikely to read Terms of Service or Privacy Policies, so they may not understand how their data will be stored or used.

The Internet Watch Foundation reported that self-generated child sexual abuse material circulating online rose by 417% between 2019 and 2022. The term 'self-generated' is imperfect, since in most cases abusers coerce children into creating such images. With AI undressing tools, children may unknowingly create AI-generated CSAM. If they upload a clothed photo of themselves or another child, someone could 'undress' it and distribute it more widely. The Internet Watch Foundation found that 99.6% of the AI-generated CSAM it examined depicted girls.

As with other deepfakes, people can use AI undressing tools to bully others. This may involve falsely claiming that a classmate sent a nude image of themselves, or creating a fabricated nude with features that bullies then mock. Sharing images of peers in this way is illegal and abusive, and it can cause profound distress to the person targeted.

Research shows that the use of these AI tools is growing, particularly to target female victims. One undressing AI site stated that its technology 'was not intended to be used with male subjects', reflecting the fact that most such tools are trained on images of women. Graphika reported a 2000% increase in spam referral links for AI undressing services in 2023, and found that 34 of these providers received more than 24 million unique visitors to their websites in one month. Experts predict further online harm, including sextortion and CSAM.

Until recently, creating sexually explicit deepfake images of adults was not a crime unless the images depicted children. The Ministry of Justice has since announced a new law that will make creating such images of adults without consent punishable by prosecution, with offenders facing an unlimited fine. Sharing intimate AI-generated images without consent was already banned under the Online Safety Act in January 2024. However, the law relies on proving intent to cause harm, which can make prosecutions difficult.

Parents and caregivers can take several steps. First, have early and open conversations about appropriate content, positive relationships and healthy behaviour, since over a quarter of UK children say they saw pornography by age 11. Second, block websites and set content restrictions on mobile networks, broadband, devices and apps to reduce the chance of stumbling on harmful material. Third, build children's digital resilience: help them recognise potential harm online, know how to report and block, think critically about the content they encounter, and understand when to ask a trusted adult for help.

Parents who are worried about their child using AI undressing tools or becoming a victim can access expert guidance and ongoing support. Useful resources include guidance on deepfakes, advice centres for online pornography and inappropriate content, safe use of artificial intelligence, and step-by-step parental control guides. Getting the right guidance is the first step toward protecting a child's online safety, and personalised toolkits can help families build a plan that fits their needs.