How do I talk to my child about AI chatbots?

The best way to talk to your child about AI chatbots is to start with curiosity, not worry: ask which ones they use and what they like about them. Then cover four things over time: a chatbot isn’t a person and doesn’t really know them; it can be confidently wrong; personal details stay private; and for feelings, health or anything worrying, they come to you or another trusted adult. Agree a few family rules, especially for homework, and keep the conversation going rather than having one big talk.

You don’t need to understand how the technology works to do this well. The advice below comes from the NSPCC, Internet Matters and Webwise, all of which have published guidance for parents on AI.

Why it’s worth talking about now

Chatbots are tools like ChatGPT, Snapchat’s My AI and Character AI that answer in chatty, human-sounding sentences. They’re also increasingly built into apps, games and homework tools children already use (NSPCC).

Many children are already using them. In Internet Matters’ research, two-thirds of 9 to 17-year-olds had used an AI chatbot, including 58% of 9 to 12-year-olds, even though many chatbots set a minimum age of 13. The same research found children use them for schoolwork, for fun and, for some, for advice about friendships and feelings.

None of that is a reason to panic. Internet Matters also points to real benefits: help learning at their own pace, practising a language and having tricky ideas explained more simply (Offaly Express). The aim is to help your child use it well, not to make them feel they’ve done something wrong.

How to start the conversation

The NSPCC’s first tip is to start with curiosity, not concern (NSPCC via Vodafone). Children clam up when a chat feels like an interrogation. Try:

• “Which AI apps do people in your class use? What for?”
• “Can you show me how you’d use it for homework?”
• “What’s the funniest or weirdest thing it’s ever said to you?”

Share how you use AI yourself too, if you do. Internet Matters suggests exploring it together: sit down and try a chatbot side by side, which lets you model good habits and set expectations early (Offaly Express).

Short, regular chats work better than one big talk. A car journey or washing up together is often easier than sitting face to face.

The four things to get across

1. A chatbot isn’t a person. It can sound warm and personal, but it doesn’t know your child and doesn’t have real feelings. The NSPCC says chatbots shouldn’t replace people in a child’s life. This matters: in a Vodafone poll of 11 to 16-year-olds, 31% of those who had used a chatbot said they’d describe it as a friend (Offaly Express). For younger children, “it’s a clever computer program that guesses what words come next” is enough.

2. It can be wrong. Chatbots make mistakes and can sound sure of themselves while doing it. The NSPCC suggests checking anything important, especially for schoolwork or health, against a trusted source. Questions you can ask aloud together (NSPCC):

• “Does this answer make sense?”
• “Where did this information come from?”
• “How could we check it?”

3. Keep personal details private. No full names, home address, school, phone number or photos. The NSPCC points out that chats may be stored or used to train the AI, so it’s worth treating a chatbot like a public space (NSPCC via Vodafone).

4. For feelings and worries, talk to a person. The NSPCC is clear that AI tools aren’t trusted adults. For health, relationships or anything upsetting, your child should come to you, another adult they trust, or a children’s helpline. In Ireland, Childline is on 1800 66 66 66 (Webwise); in the UK, Childline is on 0800 1111 (NSPCC).

Agree some family rules

Once you’ve talked, agree a few simple rules together. Children are more likely to stick to rules they helped make. The NSPCC suggests covering when and how AI can be used, including whether any topics are off-limits, and making clear that for homework it should help them understand, not do the work for them (NSPCC via Vodafone). It’s worth checking what their school’s rules are too.

The NSPCC also says using AI to make fake images or messages about other people, or to bully, is never okay. That’s worth saying plainly, especially with older children.

If you already have a family screen time agreement, add a line or two about AI. Our guide to making a family screen time agreement that sticks shows how.

Signs to keep an eye on

Most children use chatbots in ordinary, harmless ways. Internet Matters suggests gently paying attention if your child seems to be forming an emotional bond with a chatbot, sharing personal information with it, or turning to it instead of a trusted adult (Offaly Express). Its research also encourages parents to help children keep up friendships and activities offline, since a chatbot is always available to talk (Internet Matters).

If you notice any of these, don’t confiscate the phone in a panic. Ask about it calmly, the same way you started. If your child seems low or anxious generally, talk to your GP, who can point you to the right support.

What to do next

• Ask one curious question this week, such as “Which AI apps do your friends use?”, and let your child do the talking.
• Try a chatbot together for ten minutes on a homework question, and check its answer against a book or trusted website as you go.
• Agree two or three AI rules as a family: what it’s okay to use it for, what stays private, and who to talk to when something’s worrying.
• Download Webwise’s free Parents’ Guide to Generative AI, available in English and Irish, for explainers on tools like ChatGPT and My AI.