Survey Overview
The study surveyed roughly 11,000 children across 11 nations, providing a nationally representative snapshot of concerns about AI‑generated sexual imagery.
Scale of AI‑Generated Abuse
Key findings include:
- In several countries, up to two‑thirds of respondents fear AI could be used to create fake sexual images or videos of them.
- A Center for Countering Digital Hate report estimated that the AI model Grok produced 23,338 sexualised child images in just 11 days (Dec 29 – Jan 9).
- The UK’s Internet Watch Foundation flagged nearly 14,000 suspected AI‑generated images on a single dark‑web forum within one month, with about one‑third confirmed as criminal.
- South Korean authorities recorded a tenfold increase in AI‑ and deep‑fake‑linked sexual offences between 2022 and 2024, most suspects being teenagers.
Government and NGO Responses
Both UNICEF and the Center for Countering Digital Hate are urging immediate action:
- Expand legal definitions of child sexual abuse material (CSAM) to cover AI‑generated content.
- Criminalise the creation, procurement, possession, and distribution of such material.
- Require AI developers to adopt safety‑by‑design principles and to prevent circulation of illegal content.
Calls for Regulation and Safety Measures
The brief recommends that states mandate:
- Child‑rights due diligence, including child‑rights impact assessments for AI systems.
- Pre‑release safety testing for open‑source models.
- Embedding safety mechanisms across the entire AI value chain.
International Actions and Investigations
Regulatory steps already underway include:
- The European Commission’s formal investigation into whether X failed to stop Grok from generating illegal content.
- Bans on Grok in the Philippines, Indonesia, and Malaysia.
- Ongoing investigations by regulators in the UK and Australia.