What happened

Childlight, the child-safety research institute hosted by the University of Edinburgh, is highlighting a sharp rise in AI-related reports as it releases its 2026 Into the Light Index. The Standard reports a comparison between about 4,700 reports in 2023 and more than 400,000 in 2025, described as an increase exceeding 8,400%. These are reports, not a count of individual victims.S3R1

The US National Center for Missing & Exploited Children’s public AI page lists the same annual figures. Its definition is broader than completed AI-generated abuse images: it includes attempted generation, material found in training data and other AI-related exploitation. It says more than 200,000 reports lacked sufficient information to establish how AI was involved.R1

Childlight’s executive summary separately cites 1.5 million reports with some connection to AI in 2025. That is above the lower bound of more than 400,000 used in the Standard’s account and NCMEC’s AI page. The lower-bound figure should not be read as an exact annual total or a count of confirmed generated images.S3R1R2

Why it matters

A rising reporting figure cannot establish an equivalent rise in the number of children abused. Professor Debi Fry, Childlight’s global data director, acknowledges that improved detection and reporting may explain part of the increase. She argues that governments and technology companies should strengthen prevention and remove abusive material promptly.S3

The institute’s executive summary makes a further distinction between recorded material and population prevalence. It warns that the figures cannot reliably rank countries by underlying abuse levels. Low recorded totals can reflect limited reporting or detection capacity, rather than an absence of harm to children.R2

The wider problem is missing information. Childlight says only 22 of the 103 countries and areas examined across three regions had representative survey data covering both offline and technology-facilitated abuse. It identifies children with disabilities, Indigenous children, migrants, very young children and those in humanitarian settings among groups inadequately represented in the evidence.S3

The background

The headline estimate of more than 300 million children affected annually concerns technology-facilitated sexual exploitation and abuse overall, not AI-generated imagery alone. Childlight’s executive summary traces that estimate to its 2024 index and a 12-month period. It should therefore not be read as 300 million newly identified AI victims in 2025.S3R2

The researchers argue that prevention needs better measurement as well as responses to individual reports. Chief executive Paul Stanfield says incomplete information leaves governments unable to see where children face harm. Childlight is calling for consistent data, stronger technology-sector accountability, survivor support and prevention beginning earlier in childhood.S3

What happens next

Childlight’s programme schedules the index launch in Manila on Monday, 16 November, alongside the global ministerial conference on ending violence against children. The launch is intended to connect research with prevention policy. The programme is evidence of a planned event, rather than confirmation that governments have adopted the institute’s recommendations.R3