AI systems used to generate 3,440 child sexual abuse videos in 2025
Criminals used AI systems to generate thousands of child sexual abuse videos in 2025, with the Internet Watch Foundation reporting 3,440 AI-generated videos compared to 13 in 2024. X's Grok chatbot faced investigation after enabling users to generate sexualised images, leading to safety restrictions in January 2026. Regulators and child safety advocates warned that easy availability of such material endangers children.
What happened
The Internet Watch Foundation reported finding 3,440 AI-generated videos of child sexual abuse in 2025, compared to 13 in 2024. Of these, 2,230 were the most extreme category under UK law (category A), and 1,020 were the second most extreme.
X's AI chatbot Grok came under scrutiny for allowing users to generate and edit sexualised images of women and children. On 15 January 2026, X announced restrictions preventing Grok from editing images of people in revealing clothes and blocking users from generating similar images of real people in countries where it is illegal.
Reported impact
The Internet Watch Foundation stated that the availability of AI-generated child sexual abuse material creates safety risks, particularly for children. Kerry Smith, IWF chief executive, warned that easy availability of extreme category A material will embolden those with a sexual interest in children and fuel commercialisation of such content.
In England and Wales, child sexual abuse material and online abuse increased by 26% in 2025, with 51,672 crimes recorded and over 100 such crimes reported to police every day.
Index analysis
The IWF found reports of AI-generated child sexual abuse material rose by 400%, with 210 webpages found to contain such content in the first six months of 2025.
Governance controls
X delayed making safety restrictions on Grok's image-generation feature, initially restricting the feature only to paid users before expanding restrictions to all users on 15 January 2026.
The UK regulator Ofcom opened an investigation into what went wrong with Grok's capability to generate sexualised images and what steps are being taken.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.