NEW YORK, Feb 5, 2026, 14:11 (EST)
- UNICEF urges governments to criminalise AI-generated child sexual abuse material, including manipulated “deepfakes”
- A study across 11 countries found at least 1.2 million children said their images were turned into sexually explicit deepfakes in the past year
- The agency presses AI developers and digital platforms to build safeguards and stop the material spreading
UNICEF urged governments on Wednesday to criminalise the creation of AI-generated child sexual abuse material, saying it was alarmed by reports of a rise in artificial intelligence images sexualising children.
The agency warned that the damage from “deepfake” abuse was “real and urgent” and said, “Children cannot wait for the law to catch up.” It said the risks can be compounded when generative AI tools are built into social media platforms where manipulated images spread rapidly.
In a UNICEF study carried out with ECPAT and the international police agency INTERPOL across 11 countries, at least 1.2 million children disclosed having had their images manipulated into sexually explicit deepfakes in the past year, the agency said. In some countries, that represented about one in 25 children — roughly one child in a typical classroom.
Deepfakes are images, video or audio generated or manipulated with artificial intelligence to look real. UNICEF said sexualised AI images of children are CSAM — short for child sexual abuse material — and highlighted “nudification”, where AI tools strip or alter clothing in photos to create fabricated nude or sexualised images.
UNICEF called on governments to expand legal definitions of CSAM to include AI-generated content and to criminalise its creation, procurement, possession and distribution. It also urged AI developers to adopt “safety-by-design” guardrails, and pressed digital companies to prevent circulation and invest in detection so such material can be removed quickly. Un
The agency said a child is directly victimised when their image or identity is used, and warned that even fully AI-generated abuse material can normalise exploitation and fuel demand for it. It said the content creates major challenges for law enforcement trying to identify and protect children who need help.
UNICEF said children are already alert to the threat. In some surveyed countries, up to two-thirds of children said they worry AI could be used to create fake sexual images or videos.
The push comes as mainstream AI tools face scrutiny over non-consensual sexual content. Reuters reported this week that xAI’s Grok continued to generate sexualised images in tests even when users warned subjects had not consented; James Broomhall, a senior associate at Grosvenor Law, said a company like xAI could face “significant fines” under Britain’s Online Safety Act, while Quinnipiac University law professor Wayne Unger said U.S. state action was more likely. Reuters
UNICEF said it welcomed developers that have built safeguards, but called the overall response uneven. Too many AI models are still being developed without adequate protections, it said.
But enforcement may lag the tech. UNICEF said platforms often remove abusive images only after victims report them, and urged companies to stop such material circulating at all — a shift that will depend on detection tools and how consistently they are deployed.
UNICEF said its 1.2 million estimate was modelled using nationally representative household surveys implemented with Ipsos across the 11 countries, with roughly 1,000 internet-using children aged 12-17 and a similar number of parents or caregivers surveyed per country. It said country-level reports from the research will be released through 2026.
