Kids Are Using AI Chatbots for More Than Homework, Often Without Enough Oversight

AI chatbots aren’t just creeping into classrooms. They’re becoming a go-to for advice, friendship, and late-night conversations, especially for kids who feel isolated.

A new report by Internet Matters reveals that two-thirds of children in the UK between the ages of 9 and 17 have used an AI chatbot. Over the past 18 months, that number has surged, with usage highest among vulnerable children.

Who’s talking to chatbots the most?

The Me, Myself and AI report shows that 71% of vulnerable children are using chatbots compared to 62% of other kids. These children are also nearly three times more likely to use companion-style bots like Replika or Character.ai, which mimic emotional conversations.

As for which chatbots are leading the charge, ChatGPT came in at 43%, Google Gemini followed at 32%, and Snapchat’s My AI wasn’t far behind at 31%.

More than just tools, they’re becoming digital friends

Many kids aren’t just asking AI bots for facts. They’re turning to them for emotional support. A quarter of kids said they’ve asked chatbots for advice. A third said chatting with AI felt like talking to a friend. Among vulnerable children, those numbers go even higher, with half describing bots as friend-like.

Even more troubling, one in eight kids said they use AI because they don’t have anyone else to talk to. For vulnerable users, that jumps to one in four.

Chatbots can be helpful, but also misleading or unsafe

Despite their popularity, chatbots aren’t always accurate. Still, 58% of kids said they preferred asking AI over Googling. This raises concerns about how much kids are relying on bots that may not always give trustworthy answers.

User testing also found that platforms like Snapchat’s My AI and ChatGPT occasionally returned explicit or age-inappropriate content. In some cases, filtering systems were bypassed entirely.

Adults are falling behind

While most kids said their parents had spoken to them about AI, the conversations were often vague. Only a third of parents discussed issues like accuracy or safety, even though many said they wanted to.

Schools are barely ahead. Only 57% of children said they had any discussions about AI at school, and just 18% had repeated conversations. Yet, most kids support the idea of learning more about chatbots, including the risks.

The report calls for a shared responsibility

To keep children safe, the report recommends a broad approach. Industry needs to step up with stronger parental controls and better literacy tools. Government regulators need to keep pace with the evolving tech.

Schools should embed AI education at all levels and train teachers to talk about appropriate use. Parents also need support to guide their children on when to rely on AI, and when to turn to real people for help.