AI is giving offenders new ways to win children’s trust, fabricate abusive images and automate exploitation, senior UN officials warned on November 19, 2026. Speaking at a high-level General Assembly event in New York, they called for mandatory child-safety standards rather than relying on technology companies’ voluntary efforts to keep pace.
Najat Maalla M’jid, the UN secretary-general’s special representative on violence against children, said protections were repeating the failure to keep up with social media. Deputy Secretary-General Amina Mohammed argued that AI systems were reaching children without the safety scrutiny expected of children’s toys.
Abuse beyond fabricated images
The warnings extended beyond the production of sexual images. Maalla M’jid described tools that can help offenders build relationships with children, collect information about them and identify those who may be vulnerable. She outlined several ways those capabilities are being used:
- Chatbots can gain children’s trust, gather personal information and coerce them into sharing explicit images, Maalla M’jid said.
- So-called nudification apps create fabricated sexual images that can be used to humiliate, threaten or exploit a child.
- Profiling and predictive targeting can make it easier for predators, traffickers and criminal groups to reach vulnerable children.
Kay Chau of the Tech Coalition, a nonprofit alliance focused on child protection, said perpetrators may also use chatbots for abusive role-play or to generate grooming scripts. She warned of additional risks from agentic AI—systems that can carry out tasks independently.
John Trenary, who has worked on child-exploitation investigations for more than 25 years, emphasized the combination of automation, scale and wholly synthetic content. His concern was the speed of adaptation: as investigators learn to address one misuse, another emerges.
What the figures measure
Mohammed cited research estimating that at least 1.2 million children across 11 countries had their images manipulated into sexually explicit deepfakes. These are AI-generated or altered images designed to look real. The estimate concerns experiences during the year preceding the research, not a new count collected for Thursday’s event.
UNICEF published that estimate on February 4, drawing on research with ECPAT International and INTERPOL. Nationally representative household surveys covered approximately 1,000 internet-using children aged 12–17 and 1,000 parents or caregivers per country. Researchers modeled the total using national prevalence estimates, population data and estimated child internet-use rates; 1.2 million was not the number of children interviewed.
Maalla M’jid separately cited the US National Center for Missing and Exploited Children, saying reports of abuse material linked to generative AI rose 1,325% between 2023 and 2024. That figure measures reports, rather than individual children affected. She also warned that most AI-facilitated crimes against children remain unreported and under-prosecuted.
We cannot leave their safety to the goodwill of technology companies.
Amina Mohammed, UN deputy secretary-general
Safety before children get access
Maalla M’jid called voluntary industry approaches insufficient and urged global AI governance standards, including mandatory child safety and privacy by design. Those demands put safeguards in the development of systems, rather than leaving protection solely to responses after abuse occurs.
The demands build on the AI Child Safety Pledge launched by Secretary-General António Guterres in July. Its aims include making AI systems certifiably safe before they reach children and establishing zero tolerance for AI-generated child sexual abuse material.
The officials’ argument was not that AI has no role in child protection. Mohammed also identified three uses for the technology on the protective side: detecting abuse material, identifying victims and spotting early signs of grooming.
Reader comments
Newest comments first. Replies stay oldest first.