AI child abuse images surge past last year's total, UK watchdog warns
The Internet Watch Foundation found 6,310 AI-generated child abuse images in just six months — 40% above all of 2025 — and wants binding UK law to stop it.
What happened: The UK's Internet Watch Foundation, which hunts down child sexual abuse material online, says it has already assessed 6,310 AI-generated images meeting the legal definition of child sexual abuse material in the first half of 2026. That's roughly 40% more than the 4,500-plus images found across all of 2025, meaning this year's total passed last year's before 2026 was even over. Most of the images showed girls, many aged seven to 13. The IWF separately flagged a surge in AI-made abuse videos, though these latest figures cover still images only. Report Remove, a service that helps get explicit images taken offline, logged 420 reports from children who believed photos of themselves had been faked into explicit images, already above its 397 reports for all of 2025.
Why it matters: Each flagged image represents a real child, often with their actual face lifted from an ordinary photo and altered into abuse material. The IWF says AI has made this cheaper, faster and harder to trace than producing traditional abuse material, which is why case numbers keep climbing even as awareness grows. It also means school pictures, holiday snaps and everyday social media posts can become source material for abuse without a child or parent ever finding out. Campaigners argue the scale has now outgrown what voluntary industry effort and existing enforcement alone can contain.
How it works, plainly: Much of this material comes from feeding an AI image generator, sometimes a general tool, sometimes one deliberately altered, with ordinary photos of a real child and prompting it to produce explicit versions. UK law already criminalizes this at every stage: making AI abuse imagery is illegal whether or not a real photo was used, adapting a model to produce it is illegal, and sharing such a model with others is illegal too. Building a model specifically designed to generate hyper-realistic abuse imagery carries a prison sentence of up to five years. The IWF argues that despite these laws, platforms and model-makers still aren't doing enough to stop their tools being misused in the first place.
What's being done: The IWF and the National Crime Agency are urging parents to limit how freely children's photos circulate online, setting accounts to private or sharing images only within small, trusted groups. IWF policy head Hannah Swirsky wants binding legislation forcing AI companies to design models that resist misuse upfront, rather than leaning on criminal law after the harm is done. The government says existing rules already ban abuse material and the models built to create it, and that any new law must prove it would make people safer before being introduced. AI minister Kanishka Narayan has said "nothing is off the table," but no dedicated AI safety bill has been announced.