UK Tells Parents to Hide Kids’ Photos as AI Abuse Images Surge

Two clothed teenage selfies became extreme pornographic videos after predators fed them into an AI image generator, no contact with the victim required. The images surfaced through Report Remove, a confidential UK service for children whose explicit photos spread without consent. They are the examples Britain’s National Crime Agency (NCA) and the Internet Watch Foundation (IWF) now cite in new guidance urging parents to rethink posting photos of their kids at all.

That guidance asks families to lock down what strangers can see. It does not solve the harder problem underneath: investigators say they increasingly cannot tell a manufactured abuse image from a real one, and nobody has a fix for that yet.

The Three Checks Behind Britain’s New Guidance

The guidance, issued jointly by the NCA and IWF this month, does not ask parents to stop sharing pictures of their children. It asks for three specific checks before anyone hits post.

• Privacy settings – review who can see photos across every account tied to a child, not just the newest one.
• Audience – consider a ‘close friends’ group instead of a public or semi-public follower list.
• Consent conversations – revisit permissions given to schools and sports clubs, some signed before anyone anticipated AI image manipulation.

Tim Wright of the NCA said offenders are moving faster than most families realise. Images shared online, even everyday family photos, can be stolen and manipulated to create sexualised content without a child or parent knowing, he said.

Lorna Sinclair, the NCA’s child sexual abuse education manager, told the Guardian that most families have no idea the risk exists. “The average parent or carer doesn’t post a photo of their child online thinking it could be downloaded and turned into child sexual abuse material,” she said. “There are a lot of parents and carers who don’t know this problem exists.”

A Record Year for AI-Generated Abuse Imagery

The numbers behind the campaign are stark. The IWF assessed 8,029 AI-generated images and videos as showing realistic child sexual abuse in 2025, a 14% rise on the year before.

Analysts trace part of that jump to how little raw material offenders now need. Fine-tuning techniques known as LoRAs can build a realistic deepfake from as few as 20 photos in about 15 minutes, according to the IWF’s own research. A single public profile picture is often enough.

Can Investigators Tell Real Abuse Images From AI Fakes?

Increasingly, no. IWF analysts can sometimes flag visual artefacts in synthetic video, but confirming an image’s true origin usually depends on something outside the picture itself. That gap is starting to reshape how offenders defend themselves once caught.

Dan Sexton, chief technology officer at the IWF, told the Guardian he feels “very uncomfortable” telling parents not to post pictures of their children publicly. He says he sees no better option because current technology offers too little protection.

“The only way we know is when we see the authors pointing to what they created,” Sexton said, describing how analysts often rely on offenders bragging about their work on dark web forums rather than any technical signature in the file itself.

That uncertainty cuts both ways. IWF researchers have documented offenders claiming that genuine evidence of contact abuse was actually AI-generated, exploiting what the organisation calls the “liars’ dividend,” where public awareness of synthetic media manufactures deniability for real crimes. Sexton put the scale of the challenge plainly: “It’s a problem that needs to be solved. But I don’t know if it can be solved at the level of accuracy that we need, because the technology is constantly changing.”

Criminal Gangs Are Already Exploiting the Gap

The threat is not theoretical. A criminal gang took pupil photos from a UK school website, used AI to generate more than 100 sexual images of the children, then tried to blackmail the school into paying to stop the images being published, the IWF has said.

Kerry Smith, the IWF’s chief executive, called the pattern disturbing. “If someone’s imagery is online, they could be easy pickings for criminals and anyone, especially children, could find themselves being targeted,” she said.

The problem extends well beyond one campaign. A Tech Transparency Project review found dozens of nudify apps still listed on Apple’s App Store and Google Play, collectively downloaded hundreds of millions of times, despite both companies banning explicit content in their policies. NBC News separately reported that the US National Center for Missing and Exploited Children received over a million reports tied to AI-generated abuse material in just nine months.

Consent fights are not confined to nudify tools, either. Meta’s Muse feature, which lets other users tag people into AI-generated photos, has drawn its own privacy backlash, a sign that the fight over who controls a child’s likeness online is widening well past any single app.

Not Everyone Agrees the Guidance Is Enough

Even inside the child protection world, opinion splits on whether asking families to change their habits actually fixes anything.

• NCA and IWF campaigners – argue practical steps now matter because most parents have no idea the risk exists.
• Dan Sexton, IWF’s chief technology officer – calls the advice a stopgap and wants AI models legally required to be secure by design, saying “people use the term ‘security by design,’ but I don’t see that here.”
• Internet Matters – argues families should not carry the burden at all, pushing instead for an outright ban on nudifying tools.

Internet Matters, a children’s online safety charity, put it bluntly in research backing a ban on nudifying tools: “The onus to protect children from deepfake sexual abuse cannot fall on schools and parents.”

Guterres Pushes a Global Child Safety Pledge

The debate reached the United Nations five days before the NCA and IWF published their guidance. Secretary-General António Guterres opened the UN’s first Global Dialogue on AI Governance in Geneva, Switzerland, on Monday with a warning that regulation has fallen behind the technology itself.

“A technology that can reshape economies, transform the world of work, sway elections and tilt the balance of security is being deployed faster than anyone, including the people building it, can keep up,” Guterres told delegates. “Innovation needs guardrails,” he said. “If AI is to be powerful, it must be governed.”

We do not let medicine reach a child until it is proven safe. We test every toy.

AI had already reached children’s learning, friendships and most private questions before anyone tested what it would do to them, Guterres told the summit, according to the UN’s published transcript of his remarks.

He used the speech to call for a global AI Child Safety Pledge built on three rules. Prove it is safe, he said, meaning no company should deploy a child-accessible system without child-specific testing and independent oversight. Zero tolerance for sexual abuse, he continued, meaning every company must detect, report and remove such content rather than wait for someone else to find it. And never leave a child in crisis alone, requiring systems to stop and connect a distressed child to a real person instead of an algorithm.

Annalena Baerbock, President of the UN General Assembly, echoed the point, citing figures showing that 99% of deepfakes are sexual in nature and 96% target women and girls.

What Changes for Families Next

Britain is not waiting for a global pledge to take shape. The government is introducing a new criminal offence to outlaw nudification tools outright, alongside powers letting the IWF and AI developers stress-test models for the ability to generate abuse imagery.

“I am introducing a new offence to ban nudification tools, so that those who profit from them or enable their use, will feel the full force of the law,” Jess Phillips, the minister for safeguarding, said when the plan was announced. A government spokesperson added that current law is already clear that AI-generated abuse material is illegal, saying: “We will not hesitate to go further to protect children.”

Other governments are moving on parallel tracks. The European Union is preparing its recommendation on children’s social media use, due within days. Sweden has already gone further, telling parents that children under 13 should not have smartphones at all.

None of that changes what Sexton says out loud: nobody yet knows how to build a test that can tell a real victim from an AI one with the accuracy this problem demands.

Frequently Asked Questions

Is AI-generated child sexual abuse material illegal in the UK?

Yes. Images and videos of child sexual abuse are illegal in the UK whether they were made with a camera or generated with AI, and the same rule applies to altered or ‘pseudo’ images. Reports can be made confidentially to police, to Child Exploitation and Online Protection (CEOP), or through the Report Remove service run with the NSPCC.

What should parents do if a child’s photo has been misused?

Report Remove offers a confidential route for under-18s to flag nude or sexual imagery for removal, and it includes mental health support alongside the takedown request. The IWF’s own advice is to stay calm, reassure the child they are not to blame, and report to police or CEOP immediately.

How do nudify apps actually work?

Stanford researchers describe nudify apps as tools trained on pornographic datasets that take an uploaded photo of a clothed person and return a realistic fake nude within seconds. Many of these apps do not consistently block uploads of images showing minors.

Can tech companies be held liable for AI-generated abuse content?

That question is now being tested in court. Families in the United States filed a class action in March 2026 alleging that xAI’s Grok tool generated sexualised deepfakes of minors, arguing the company skipped safeguards used by other major AI developers.

Is this only a UK problem?

No. UNICEF has called the global rise in AI-manipulated sexual imagery of children alarming, and US authorities have logged over a million AI-related abuse reports in a single nine-month stretch. Guterres’s push for a global Child Safety Pledge reflects a problem regulators across multiple countries are racing to catch up with at the same time.