New research reveals 40 percent increase in AI-generated child sexual abuse material reported to authorities
Guardii Analysis
The Internet Watch Foundation published its 2026 annual report on October 30, revealing a 40 percent year-on-year increase in confirmed reports of AI-generated child sexual abuse material, with analysts assessing 12,500 unique AI-generated images and videos during the reporting period compared to 8,900 the previous year. The Foundation noted that the quality and realism of synthetically generated material has improved significantly, making it harder for both automated systems and human reviewers to distinguish AI-generated content from photographs of real children, and that offenders are increasingly using generative AI tools to create bespoke material tailored to specific requests in online communities. The report also documented cases in which real children's social media photographs were manipulated using AI face-swapping and deepfake technology to produce sexualized imagery, a tactic the Foundation described as a growing threat that blurs the line between synthetic and real-victim material.
The detection challenge the Internet Watch Foundation describes, distinguishing AI-generated from real imagery and identifying manipulated content at scale, is one part of the operational picture; the other is recognizing when a child is being targeted with such material as a grooming or sextortion tool. Systems designed to detect child sexual abuse material, including AI-generated and deepfake content, in the context of a live conversation rather than in bulk image repositories address both dimensions: Guardii's anti-CSAM module identifies when sexually explicit material involving apparent minors is shared in direct messages on platforms including Instagram, Snapchat, Discord and Roblox, regardless of whether the image is a photograph, a synthetic creation or a manipulated deepfake, and escalates immediately when such material appears in a conversation with a child, enabling intervention before the imagery is used to coerce, blackmail or normalize abuse.