The Internet Watch Foundation, a UK-based group that finds child sexual abuse images and videos online and works to stop them spreading, says AI-made abuse imagery is growing fast. Its analysts assessed 6,310 photorealistic AI-made images that met the legal definition of child sexual abuse between Jan. 1 and June 30, 2026. That is 40% more than the 4,512 they found in all of 2025. The group published the figures on Monday, Oct. 5, and used them to press the European Union to agree on a law first proposed in 2022.
The images mostly show girls, and more of them show younger children. Where analysts recorded both age and sex, girls appeared in 98% of the images. Children aged 7 to 13 made up 79% of the images, up from 70% in 2025, and 190 images showed babies and toddlers under 2. Most of the images, 88%, fell into Category C, the UK legal grade that can include sexualized posing and nude images; 350 were in Category A, the most severe. The foundation says the images cause real harm: in some cases offenders use AI tools to alter real photos of victims, which can further traumatize survivors.
The way the foundation fights this material shows why it is worried. It turns confirmed criminal images into hashes, unique digital fingerprints, which police and tech companies can use to spot and block copies automatically without anyone viewing the images again. A fingerprint can only match an image that has already been found, though, and a brand-new AI-made image has none. So the foundation wants EU rules that let companies detect previously unseen material as well as known images.
The European Commission proposed permanent rules in May 2022 that would oblige online services to detect, report and remove child sexual abuse material. Negotiations between the European Parliament and EU governments began in December 2025 and are still going on, with a sixth round set for Sept. 29. No final deal has been announced. In the meantime, a temporary rule that lets companies scan for such material voluntarily lapsed in April and was brought back in July, to run until April 3, 2028. The revived rule leaves out encrypted messaging services.
"If Europe is serious about protecting children, it needs a comprehensive Child Sexual Abuse Regulation that gives platforms the legal certainty to detect, prevent and respond to known and unknown child sexual abuse content," said Kerry Smith, the foundation's chief executive.
Encryption is still an open question. EU governments said in July that accepting the Parliament's exemption for encrypted messaging in the temporary rule does not mean they would accept the same in the permanent law. It is not clear whether negotiators will allow detection of new images, or how that would work in encrypted apps. The foundation's release also gives no figures for AI-made videos and does not say which AI tools made the images.