United Nations officials warned that artificial intelligence is giving predators and criminal groups new tools to target children, and that efforts to protect them are falling behind.

“We lost the battle on social media because we were lagging behind, and we are now losing the battle on artificial intelligence because we are still lagging behind,” said Dr. Najat Maalla M’jid, the UN Special Representative of the Secretary-General on Violence against Children. She spoke Thursday at a high-level event in the General Assembly on AI’s risks and opportunities for ending violence against children at UN Headquarters in New York.

UN Deputy Secretary-General Amina Mohammed said that over the past year, no fewer than 1.2 million children across 11 countries had their photos altered into sexually explicit deepfakes, citing a recent study by the UN Children’s Fund (UNICEF). Dr. M’jid said reports of child sexual abuse material tied to generative AI increased by 1,325% between 2023 and 2024, according to the U.S. National Center for Missing and Exploited Children, adding that most AI-facilitated crimes against children remain unreported and under prosecuted.

Dr. M’jid said chatbots are being used to gain children’s trust, harvest personal information and pressure them to share explicit images, while deepfakes made with so-called nudification apps are used to humiliate, threaten and exploit them. She said profiling and predictive targeting make it easier for predators, traffickers and criminal groups to find and reach vulnerable children. John Trenary, a veteran of more than 25 years of child exploitation investigations, said predators have long sought to weaponize technology, and warned that AI’s ability to scale and automate synthetic content makes misuse self-perpetuating.

Kay Chau of the Tech Coalition said perpetrators may be using chatbots to script grooming and to facilitate abusive role-play, and that agentic AI — systems that carry out tasks independently — is introducing new risks.

Mariana Rozo-Paz of the Datasphere Initiative described a digital literacy program with teenage girls in rural Colombia who depicted an AI agent they called Lucia as part of their online lives, saying it could act in “intimate mode” or “best friend mode.”

Mohammed said AI systems are reaching children without the kind of safety checks applied to physical products. “We would never accept a children’s toy being put on the market without proper safety checks, but AI systems are reaching children without the same scrutiny,” she said. She added that safety cannot be left to the goodwill of technology companies.

The UN said Secretary-General António Guterres launched the AI Child Safety Pledge in July. According to the UN, its goals include ensuring AI systems are certified safe before children encounter them, and zero tolerance for AI-generated child sexual abuse material. Dr. M’jid said voluntary industry approaches are insufficient and called for worldwide rules to anchor AI governance, with required child safety and privacy built in from the start.

Mohammed said AI can also be used to find abusive content, help identify victims and flag early grooming behavior. In a World Childhood Foundation study cited by the UN, a young person in Sweden said technology is advancing faster than those tasked with controlling it, and the foundation’s Secretary General, Paula Guillet de Monthoux, said adults must show they can keep up.