Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

The Internet Watch Foundation (IWF) identified 3,443 AI-generated child sexual abuse videos in 2025, up from 13 in 2024—more than 260 times as many. That figure counts videos, not all AI-generated material, and it is an operational detection total, not an estimate of how much such content exists worldwide.

What the IWF’s annual video count shows

The IWF’s 2025 Annual Data & Insights Report says it identified 3,443 AI-generated child sexual abuse videos in 2025, compared with 13 in 2024. The increase is more than 260-fold. These are videos identified by the IWF, not a count of every video produced or circulating online.

Category A made up a larger share of the AI-generated videos

The IWF classified 65% of the AI-generated videos—2,233—as Category A. For comparison, 43% of videos involving real children were Category A. This compares the category shares of two video groups; it does not mean that 65% of all AI-generated material falls into Category A.

Why another IWF figure is higher

In a March 2026 announcement accompanying Harm without limits: AI child sexual abuse material through the eyes of our Analysts, the IWF reported identifying 8,029 realistic AI-generated images and videos in 2025, 14% more than in 2024. It also reported 82 prohibited items that were not photorealistic. The announcement describes separate measures: the 8,029 figure combines images and videos, while the 82 items are reported separately. Neither should be added to the annual video count as though they were the same unit.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

What the earlier 2025 snapshot counted

A July 2025 IWF update covered the first half of that year, not the full year. It reported AI-generated material on 210 webpages from January through June 2025, compared with 42 webpages in the same period of 2024. It also reported 1,286 AI-generated videos in the first half of 2025, compared with two in the equivalent 2024 period. The mid-year figures count webpages and videos separately; a webpage can contain multiple items, so the webpage count is not a file count.

Why these figures are not a global prevalence estimate

The IWF says its annual findings arise from its operational work, including proactive detection, and should not be interpreted as a measure of global prevalence. The totals show what the organisation identified under its work and methods; they do not establish the total amount of material online, the number of people making it, or the proportion of all online abuse represented by AI-generated content. The IWF makes this limitation explicit.

Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Support on Ko-Fi

How AI changes the risks described by the IWF

The IWF says generative AI tools—including video models, nudification apps, subscription platforms and agentic AI systems—can lower technical barriers to producing and distributing abusive material. Its analysts also describe synthetic content, the manipulation of real children’s imagery, and simulated explicit chats. The harms may extend beyond the creation or circulation of synthetic content: using a real child’s image can re-victimise that child, while manipulated imagery or chats may be used in blackmail or sexually motivated extortion. The IWF outlines these risks.

IWF Chief Executive Kerry Smith said on 16 January 2026: “When images and videos of children suffering sexual abuse are distributed online, it makes everyone, especially those children, less safe.” The statement underscores that distribution itself can cause harm, including when imagery is synthetic or manipulates a real child’s image.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.