Two in three UK teenagers turning to AI chatbots for advice and companionship
Key Points
- A Youth Justice Board evidence pack published in June 2026 reports that 67 per cent of UK children aged 15 to 17 use AI chatbots for advice, support or companionship.
- The pack warns that AI chatbots may give children inaccurate or harmful information and reinforce problematic behaviours.
- The National Crime Agency warns that generative AI is heightening risks of harm, particularly in relation to child sexual abuse.
- Online child sexual abuse accounted for at least 42 per cent of recorded offences of this kind in England and Wales in 2024, up from 35 per cent in 2023.
- Ofcom opened a formal investigation into the platform X in January 2026 over AI-generated abuse content.
Around two in three teenagers in the UK are turning to AI chatbots for advice, emotional support or companionship, according to a new evidence pack published by the Youth Justice Board (YJB).
The pack, released in June 2026, cites research estimating that 67 per cent of children aged 15 to 17 in the UK use AI chatbots, which have evolved to closely mimic human interaction.
It warns that the technology may provide children with inaccurate or harmful information, reinforce problematic behaviours and encourage violence towards others.
The findings reference several lawsuits filed in the United States against the chatbot platform Character.ai, which allege that the service encouraged children to harm themselves and others.
The evidence pack draws on warnings from the National Crime Agency (NCA), which states that generative AI is heightening existing risks by broadening both the extent and impact of harm.
The agency cautions that the technology is increasingly being used in serious organised crime, particularly in relation to child sexual abuse.
In 2024, online child sexual abuse and exploitation accounted for at least 42% of all recorded offences of this kind in England and Wales, up from 35% in 2023.
The pack states that AI-generated imagery is contributing to the increase in child sexual abuse material online, with such images becoming increasingly realistic.
It notes that generative AI tools available on mainstream social media platforms have been misused to produce and share abusive material.
In January 2026, Ofcom launched a formal investigation into the platform X under the Online Safety Act following such concerns.
The pack also identifies wider risks associated with the technology, including its potential use to evade content moderation, tailor extremist propaganda, and create chatbots designed to lead users, including children, towards more extreme content.
The warnings form part of a broader assessment of online harms affecting children, which the YJB says require coordinated action across youth justice, health, education, technology companies and regulators.
In June 2026, the government instructed technology companies to introduce device-level controls to prevent children from seeing, sending and receiving explicit content, warning that legislation would follow if firms failed to act within three months.
