Eigen RadarAI
Archive
Analysis

IWF’s six-month AI abuse-image count exceeds the whole of 2025

IWF’s analysts counted 6,310 still images depicting child sexual abuse and made with AI during January–June 2026. That six-month assessment already surpasses the organisation’s total for the previous year. A single Guardian report sets out the new figures and IWF’s demand for binding legislation. The dataset measures assessed material, rather than every abusive image available online.

Artificial Intelligence··Morning
An adult in a teal blouse, seen from behind, works at a computer beside closed case folders.

Six months of assessed images exceed the previous year’s total

IWF, the Internet Watch Foundation, is a UK-based organisation monitoring child sexual abuse material globally. Its analysts identified 6,310 AI-made still images depicting legally defined abuse during the opening six months of 2026. The previous year’s assessment total was above 4,500; the shorter new period has already passed it. These figures come from IWF through a single Guardian report and concern still images, not videos.[1]

The assessed material predominantly showed girls, with children between seven and thirteen particularly represented. In January–June, analysts also found a greater volume of photorealistic material than across the preceding year. The counts concern what the organisation examined, rather than measuring everything in circulation online.[1]

The new figures strengthen IWF’s call for binding legislation

IWF policy head Hannah Swirsky backs calls from parliamentarians for enforceable AI rules that would oblige developers to make their models safer. No imminent AI-specific bill has been signalled by the government. Kanishka Narayan, the UK minister responsible for AI, leaves the possibility of further regulation open, saying legislative proposals should be assessed against whether they improve safety for people in Britain.[1]

Existing UK law also covers AI-generated abuse material

UK law already prohibits child sexual abuse material made with AI, as well as altering models to generate it or distributing such adapted models. Anyone developing a model specifically for highly realistic abuse imagery faces imprisonment of up to five years. Report Remove, which helps stop intimate pictures appearing online, has separately received 420 reports from children who believed explicit images of themselves were fabricated or altered. That exceeds the service’s 397 reports during 2025.[1]

References

  1. News sourceThe GuardianIWF identifies more AI abuse images in six months than in all of 2025↩1↩2↩3↩4
AI safety and regulationChip wars and the supply chain

More analysis from this desk

View all Brief analyses from the Artificial Intelligence desk →