On October 5, the UK's Internet Watch Foundation (IWF) released new figures: between January 1 and June 30, 2026, its analysts confirmed 6,310 AI-generated child sexual abuse images — 40% more than the 4,512 confirmed for the whole of 2025. Six months' worth of material already outstrips the entire previous year.
The IWF is a UK nonprofit dedicated to finding and getting child sexual abuse material removed from the internet. Its count only includes images that analysts have individually reviewed and confirmed meet the UK's legal definition.
A victim profile skewing younger
Among the 6,221 images where analysts recorded both age and gender, girls appeared in 6,094 — 98% of the total. For all of 2025, 4,259 images depicted girls.
By age, children aged 7 to 10 accounted for 2,534 images and those aged 11 to 13 for 2,369, together making up 79% of the total, up from 70% across 2025. Children aged 3 to 6 appeared in 1,004 images, and those under 2 in 190.
Under the UK's severity classification, Category C — the least severe — accounted for 5,557 images, about 88%; Category B for 403; and Category A, the most severe, for 350.
Doing the monthly math
The 40% figure compares half a year against a full year, so the gap looks even starker on a monthly basis. The first half's 6,310 images work out to roughly 1,050 a month; 2025's 4,512 average to about 376 a month. On that rough calculation, this year's monthly discovery rate is running at close to 2.8 times last year's.
That comparison has limits. How much material gets found depends on how many analysts the IWF puts on the problem and how many tips come in, as well as how fast the generation tools themselves are spreading. The IWF's figures don't separate those factors out, so there's no way to say how much of the increase reflects more material actually existing versus better detection.
Video is a separate line of tracking. This release covers only images, but the IWF has previously reported finding 3,440 AI-generated child sexual abuse videos in 2025, up from just 13 in 2024.
The ask lands in Brussels
The IWF is directing its message at EU lawmakers. The European Commission proposed a draft Child Sexual Abuse Regulation in 2022, and the IWF is still pushing for the bloc to pass a comprehensive version of it.
The IWF's argument is a technical one. Platforms currently catch this material mainly through hash matching — comparing the digital fingerprint of a newly uploaded image against a database of known samples, and blocking it if there's a match. Because AI generates a fresh image every time, there's no matching sample in any database, so hash matching comes up empty. That's why the IWF wants the regulation to cover detection of both "known" and "unknown" content.
IWF Chief Executive Kerry Smith said in a statement:
"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."
The draft has been debated inside the EU for years. Opponents are mainly worried that requiring platforms to proactively scan private messages — potentially including encrypted communications — would undermine the privacy of communications. Detecting "unknown" content is exactly what requires that kind of proactive scanning, which puts the IWF's ask right at the center of the dispute.
The numbers still missing
The IWF didn't disclose which tools generated the images in this release — how much came from open-source models, commercial services, or modified local deployments isn't broken out publicly. Whether the growth rates in both images and video will hold through the second half will depend on the IWF's full-year figures.
Sources: IWF statement, The Print, CocoLoop, Irish Examiner; the IWF statement is the source for the image counts, gender/age breakdown and severity classification, while the video figures come from the IWF's 2025 annual report.