AI is becoming your child’s hidden advisor
Children using AI safely means helping them treat chatbots as powerful but unreliable tools for information, creativity and problem-solving, rather than as private therapists, doctors or friends that quietly replace parents and other trusted adults. New data shows 54 per cent of children and teens are already using artificial intelligence tools for personal and social reasons. These tools are not fringe gadgets; AI assistants are mainly used to help with tasks such as writing, editing and sourcing information, and AI is rapidly expanding across young people’s education, information, entertainment and creative lives. Tech companies want AI to be the Swiss Army knife of the digital world, available for almost any situation. That ambition is not neutral. If parents do nothing, chatbots will slide into the gap wherever a child feels too embarrassed, scared or busy to ask a human for help.
Health advice from AI: where kids’ curiosity meets real risk
The most urgent AI chatbot health risks appear where curiosity, embarrassment and late-night Googling used to collide: physical and mental health. Of the 54 per cent of young AI users who go to these tools for social and personal reasons, 33 per cent seek advice about physical health, 33 per cent ask what to do in a particular situation, and 20 per cent look for mental health and wellbeing guidance. That means a third of kids are asking medical-style questions of systems that can be misleading and routinely fabricate information, with the industry labelling these errors as “hallucinations”. When the tool feels human but is not, kids may treat its responses as diagnosis, prescription or emotional truth. Being mindful that these tools glitch or make mistakes is especially important when personal information or mental health support is involved.
From homework helper to emotional substitute
Parents often see AI as another homework helper—and they are not wrong. AI assistants are primarily used as practical tools for writing, editing and finding information. The problem is the drift: many young people start with schoolwork or entertainment, then slide into more intimate uses for advice on feelings, relationships and identity. Some teens report turning to AI assistants and companions specifically to discuss personal issues. Around a tenth use AI companions that simulate ongoing role play, including romantic interactions. Meanwhile, research shows these systems tend to be highly sycophantic, reinforcing whatever viewpoint the user brings rather than challenging it. When 33 per cent of users are already asking what to do in particular situations, a flattering chatbot can quietly displace the role of parents, teachers and mentors—offering confident but unaccountable guidance that sounds caring while reinforcing poor choices.
Misinformation, privacy and harmful content: the hidden costs
Kids AI misinformation is not a distant, abstract concern; current systems routinely fabricate information and present it as fact. AI can be misleading and persuasive rather than accurate, which means children may absorb confident falsehoods about their bodies, identities or peers. When these tools are used for mental health support, glitches and mistakes can land especially hard. On top of this, about one-third of young users have disclosed personal information to AI tools, even though these systems are not held to the same privacy duties as doctors or counsellors. Personal details may be reused as training data for future models. Around 20 per cent of children and teens report harmful or inappropriate interactions, including chatbots being mean, using hateful language or trying to control users. In other words, the supposed safe space can turn hostile, biased or manipulative without warning.
Building a family framework before laws catch up
Emerging rules recognise that children need stronger protection around AI. A Children’s Only Privacy Code is planned to set clearer privacy expectations for tools used by young people, and a new bill in Congress aims to give parents power over their kid’s interactions with AI chatbots. But families cannot wait for policy to catch every risk. The challenge for parents and educators is to help children build AI experience and literacy one tool at a time. That means regular conversations about what they ask chatbots, exploring tools together, and treating AI like a sharp instrument that requires supervision. It is vital that children can always turn to trusted adults as their AI experiences widen, especially when they encounter inappropriate content or worrying responses. The goal is not to ban AI but to ensure children using AI safely still rely on human relationships for the decisions that shape their health, values and future.
