TL;DR

The UK’s Independent Anti-Slavery Commissioner, Eleanor Lyons, has told The Independent that criminal gangs are using AI to reach and exploit more children. In an interview published on 24 October, she said AI-generated profiles posing as young people are being used to groom children into carrying drugs or cash, and are hard even for experts to detect. She wants a dedicated child exploitation strategy from the government.

How AI is being used

Lyons said criminals use AI the way everyone else does, to work faster and reach more people. That includes running several fake profiles at once to build trust with a child before recruiting them. She said AI is “making it far more hard for us to understand actually what’s a real profile and what’s not”.

The tasks children are drawn into can look harmless at first, such as paying a small sum into their bank account and withdrawing it the next day. The amounts then grow, turning the child into a money mule. Some are later coerced into violence for the gangs.

A changing set of victims

In 2025, 7,045 children were referred as possible victims of modern slavery or trafficking, 17% more than the year before, and British nationals made up more than 60% of them. Lyons said the profile of those targeted is shifting. Where her concern used to centre on children in care or going missing, she now hears from families of children from stable homes and good schools. Because these children are unknown to social services, gangs see them as harder for police and councils to spot.

Pressure on tech firms

Lyons welcomed the government’s planned ban on under-16s using platforms such as Instagram and Facebook, but said more is needed. She argued that tech companies are not fully meeting their duties under the Online Safety Act, including on taking down harmful content involving children. “Our policy response isn’t keeping up,” she said, with the speed at which criminals adopt new technology.

Looking Forward

The commissioner’s call for a specific strategy comes as the under-16s ban approaches. In our view, the detection problem she describes is the harder one: if AI-generated profiles cannot be told apart from real children, platform moderation and parental vigilance both weaken. The government had not responded to The Independent at the time of publication.