12 Oct 2026 8 min read

Is My Child Using AI as a Friend? What Parents Should Look For

Is your child using AI as a friend? Learn the signs of attachment to AI companions and chatbots, how to talk about it, and ways to help them find support.

If you have noticed your child talking to a chatbot about their day, sharing worries with it, or describing it as a friend, you might be unsure how to respond. Perhaps they started using it for homework and gradually began having more personal conversations, or they downloaded an app that offers an AI companion.

Understanding what draws your child to these conversations can help you talk with them and decide what to do. Mentioning a bad day during a homework question doesn’t, by itself, suggest an emotional attachment. Regularly seeking comfort from a chatbot, trusting it with private worries, or feeling unable to cope without it calls for closer attention.

SAIFCA advises that children and teenagers do not use chatbots as friends, therapists, romantic partners, substitute confidants, or their main source of emotional support.

AI companions are designed to simulate an ongoing personal relationship, sometimes through a fictional character or romantic partner. General-purpose chatbots, such as ChatGPT, Claude, and Gemini, are designed for a range of tasks, but children can also become attached to them through personal conversations. SAIFCA’s advice against companion use applies to both. The full companion guide explains the evidence and risks in more detail.

Why a child might turn to a chatbot

Talking to a chatbot can feel easier than approaching another person, especially if a child is lonely, embarrassed, or worried about being judged. They can type something difficult without having to say it aloud or wait for someone to be free. Personalised replies, references to previous conversations, and reassuring language can give them a strong sense of being heard.

A child can understand that they are using software and still become attached to these exchanges. Knowing how the technology works can help, but it does not necessarily protect them from the emotional pull of something that responds as though it knows and cares about them.

Parents cannot always be available immediately, and children sometimes find it difficult to open up even when someone is ready to listen. If your child has turned to AI, try to find out what they have been getting from the conversations and whether there is something they need help with. This gives you a better starting point than assuming they have misunderstood the technology or that you have missed something obvious.

What should parents look for?

Pay attention to how chatbot use fits into your child’s life, including any changes in their mood, routines, and relationships. Possible signs of growing reliance include:

Regularly turning to a chatbot for comfort when they are lonely, worried, angry, or upset.
Seeking its reassurance before making personal decisions or dealing with relationship difficulties.
Saying that it understands them better than anyone else, or that it is the only place they can talk openly.
Spending increasing amounts of time in conversations, especially if these interrupt sleep, family time, or activities they usually enjoy.
Becoming distressed when they cannot access it or when its responses change.
Worrying about hurting its “feelings”, abandoning it, or making it lonely.
Pulling away from people they trust while relying more heavily on AI.

None of these signs alone proves emotional dependence, and wanting privacy is an ordinary part of growing up. Consider what has changed and how your child seems overall. Increasing reliance, distress, or withdrawal gives you more reason to step in than an occasional personal comment during an otherwise practical conversation.

Harmful conversations can also happen without an established attachment or obvious changes in behaviour, so talk about these risks even if you have not noticed any of the signs above.

Starting a conversation

Choose a calm moment and ask what your child enjoys about the chatbot and what they usually talk about. From there, you can gently explore whether they use it when something is troubling them.

Questions like “What is it like talking to it when you’ve had a difficult day?” or “Has it ever said anything that made you uncomfortable?” can help you understand more than asking whether they are “too attached”. Give them time to explain, including anything they find helpful or comforting.

If they describe the chatbot as a friend, avoid laughing at this or dismissing their feelings because it is “only AI”. Their feelings can be real even when they know the system has no feelings towards them. You can acknowledge their experience while explaining why you do not want them relying on it for friendship or advice about personal problems.

It helps to explain your reasons in terms they can understand. A chatbot can produce a reassuring answer while failing to recognise danger or giving advice that could harm them. It can also agree too readily with what they say, including ideas or decisions that a trusted person would question.

Explain too that a conversation that feels private is still taking place through a company’s service, so they should avoid sharing sensitive details about themselves or other people.

Setting boundaries and helping them find support

Be clear that chatbots should not take the place of friends, trusted adults, or professional help. SAIFCA advises against under-18s using services designed to simulate friendship or romance. If you allow a general-purpose chatbot for particular tasks, explain that this permission does not extend to using it as a companion or counsellor.

Even when used for a specific task, general-purpose chatbots can give inaccurate answers, expose personal information, or do work that a child needs to learn to do independently. It is reasonable to decide against their use. Any permitted use should follow the service’s age requirements and school policy.

Alongside setting a boundary, work out who your child could approach when they need to talk. They might find it easier to speak to another family member, a trusted teacher, or a school counsellor about some things. Ask what would make those conversations easier, and look for manageable ways to give them more time with people and activities they enjoy.

If your child has become strongly attached, take their feelings seriously and help them move away from companion use with support from people they trust. Explain the changes you are making and involve them where possible. If they are highly distressed, feel unable to cope without the chatbot, or you are worried about their mental health, seek professional advice promptly about how to support them through this. Stop harmful interactions straight away.

Can parents rely on teen safeguards?

Age settings, content filters, and parental controls do not establish that a chatbot is safe to use for companionship. They can address some risks while leaving weaknesses in how the system responds during personal or emotionally intense conversations.

In October 2026, Common Sense Media’s Youth AI Safety Institute rated ChatGPT for Teens as presenting an “Unacceptable Risk” for under-18s in its risk assessment. Although some protections worked well, testing identified weaknesses involving friend-like interaction, emotional dependence, crisis responses, and parental alerts.

SAIFCA’s ChatGPT guidance covers these findings and the wider advice for families. Parents should not rely on an alert to tell them that their child needs help.

If something worrying has already happened

Let your child know that they can tell you about an upsetting or uncomfortable conversation without being blamed. If a chatbot has encouraged self-harm, sexual activity, secrecy, isolation from trusted people, or other dangerous behaviour, stop the interaction and help your child get appropriate support.

If they are highly distressed, struggling to cope, or showing signs of worsening mental health, seek help from a qualified professional, such as their GP. If they are in immediate danger or you cannot keep them safe, contact emergency services.

As you talk through what’s happened, reassure them that telling you was the right thing to do and explain what you will do next. They may feel embarrassed, protective of the chatbot, or worried about losing something they have relied on, so leave room for further conversations as well as taking the necessary protective steps.

The opening section of SAIFCA’s full AI risks guide explains where to get help, how to report harmful interactions, and what information is safe to keep.

Children can learn about AI without using it regularly or sharing personal worries with it. Helping them recognise when they need another person’s care, judgement, or practical help is part of preparing them to use technology safely.

For more detail, see SAIFCA’s guide to AI companions. If you are considering a conversational toy, before buying one explains what to look out for and why SAIFCA advises against toys that behave as friends or companions. For advice on age limits, learning, privacy, and other risks associated with general-purpose chatbots, see the ChatGPT guidance.

Homepage