scorecardresearch

Tuesday, November 17, 2026

Support our Journalism

Depth, context and analysis

Subscribe
HomeFeaturesChild sexual abuse in AI era—US reports 8,400% surge in cases in...

Child sexual abuse in AI era—US reports 8,400% surge in cases in two years

Meta reported that it proactively took action on 5.3 million pieces of child sexual exploitation content on Facebook and Instagram in India during the first half of 2026.

Stay updated with ThePrint:

New Delhi: Reports of child sexual exploitation involving the use of generative AI have surged worldwide, while governments and tracking organisations still struggle to establish the scale of the abuse. This is despite the efforts made by tech companies, which are adopting new measures to stop the spread of AI-generated child sexual abuse material (CSAM). 

The most alarming figures emerged from the 2026 Into the Light Index released on 16 November by the Childlight Global Child Safety Institute. The index draws on 645 studies of offline abuse across 47 countries and 76 studies of technology-facilitated abuse across 21 countries. The most notable statistic found is that reports of child abuse involving the use of AI to the US National Center for Missing & Exploited Children (NCMEC) shot up from 4,700 in 2023 to over 400,000 in 2025, an increase exceeding 8,400 per cent. 

The reports to NCMEC range from AI-generated abuse material or attempts to produce such material, as well as other forms of AI-facilitated exploitation such as extortion. The researchers say the sharp increase could be at least partially due to changes in detection methods and reporting practices.

Childlight estimates that over 300 million children worldwide experience at least one form of sexual exploitation or abuse that is facilitated by technology annually. This is even though only 22 of the 103 countries and territories examined had representative surveys that covered both online and offline forms of abuse. 

“What remains invisible risks remaining unaddressed,” Childlight chief executive Paul Stanfield said in the institute’s accompanying statement when calling for improved data collection and prevention.

AI-generated abuse material increasing

A separate Internet Watch Foundation (IWF) assessment reached similar conclusions about the rise in AI-generated CSAM: it identified 6,310 AI-generated images that meet the UK’s legal definition of CSAM in the first half of 2026, compared with 4,512 in all of 2025.

In the AI-generated images for which age and sex were recorded, 98 per cent included girls, while the IWF’s broader findings indicated that girls accounted for 77 per cent of the children depicted in abuse images assessed during 2025. The figures concern detected images rather than the number of individual victims, however, and cannot be used to establish a population-wide prevalence of abuse.

Hannah Swirsky, the IWF’s head of policy, called for “binding legislation on AI” in an interview with The Guardian.

“It is incumbent on tech companies to build tools which cannot be abused this way,” she said.

What AI companies are doing

In April 2024, OpenAI, Google, Meta, Microsoft, Anthropic and other developers adopted the Safety by Design principles developed by child-protection organisation Thorn and non-profit think tank All Tech Is Human.

The commitments include, but are not limited to, screening training datasets, testing models against abusive requests, and preventing the generation or distribution of exploitative material.

The commitments have had some effect, as, according to Thorn’s 2025 impact report, the participating companies collectively managed to block hundreds of thousands of prompts to generate harmful content and removed hundreds of models designed to produce abusive imagery. However, these results were self-reported by the companies and were not independently audited.

OpenAI announced on 16 November that they made a new image classifier specifically designed to detect potential CSAM, which extends to material that hasn’t been generated yet. Internal testing of the classifier concluded that it could flag 48 per cent fewer images for human review while detecting approximately the same number of abusive images as its existing technology. OpenAI plans to offer the classifier only to vetted organisations free of charge.

On the social media side, Meta reported that it proactively took action on 5.3 million pieces of child sexual exploitation, not limited to AI-generated content, content on Facebook and Instagram in India during the first half of 2026, with more than 98 per cent identified before users reported them.


Also read: ‘Love Pattinson, hate this choice’—fans react to new James Bond casting


India faces major evidence gaps

Childlight’s profile on India identified only two small representative studies examining technology-facilitated abuse. They reported higher exposure among boys to certain forms of unwanted sexual content and online solicitation, while four other studies examining offline sexual assault consistently found higher reported victimisation among girls.

What this highlights most of all is the lack of data availability and limited national-level data, which prevent comprehensive assessments of the prevalence of abuse in India, regardless of whether it is AI-generated or not.

Despite the previously discussed corporate action, NCMEC warns that preventive measures against AI-generated abuse have neither been universally adopted nor have they been made mandatory across generative AI services, leading to inconsistent implementation and action.

Professor Debi Fry, Childlight’s Global Director of Data at the University of Edinburgh, called for wider government support and legislation: “By bringing together evidence from surveys, police records and other sources, we can build a clearer picture of the risks children face, identify weaknesses in protection systems and help target resources and interventions where they are needed most.”

(Edited by Ratan Priya)

Subscribe to our channels on YouTube, Telegram & WhatsApp

For readers who want the full picture.

You read our articles closely. Subscribe to go further, with exclusive newsletters in your inbox, flagship events hosted by Shekhar Gupta, AMAs with our senior editors, the chance to write under your own byline, and more.

Subscribe to ThePrint →

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Most Popular