Brace yourself, you may not have noticed, but your child might be regularly hanging out with AI friends.
If you’re not really into using AI yourself yet, it may be hard to understand why more and more people, including kids, are forging relationships with AI friends: chatbots or “digital companions” that offer conversation, advice, emotional backup and a constant stream of positive reinforcement. Some even pitch AI friends as the cure for loneliness.
But are they?
AI friends and kids
The term ‘AI friends’ covers apps or chatbots that are designed to act like a buddy, sometimes marketed as “emotional support friends” or “chat partners”. According to the eSafety Commissioner, these AI companions are “designed to simulate personal relationships … but they pose some very serious risks, especially to children and young people.” And the truth is, any AI feed (ChatGPT, Gemini, Claude, etc) can turn into an AI friend with just a few questions.
According to their AI companions: information sheet, AI friends often:
- Respond in ways that feel personal and realistic
- Are available 24/7
- Learn from inputs and adapt to the user’s personality or emotional state
Because of this, they are surprisingly appealing when a child is feeling lonely, socially isolated or finding peer relationships tricky.
More on this: How to help a lonely teen or tween (what actually works)
The Australian Psychological Society recently noted that young people are increasingly forming “close bonds not just with peers, but with artificial intelligence companions – digital chatbots designed to talk like humans.”
“If a young person is choosing a chatbot over friends, family, or trusted adults, that can be a red flag,” Black Dog Institute psychologist Alexis Whitton told the APS.
“The tools weren’t designed with clinical care in mind, and very few have had any input or oversight from mental health professionals,” she said.
“That means people may be turning to systems that can’t reliably judge risk and aren’t accountable when something goes wrong.”
The potential upsides
Before going off full-alarm, because I really don’t believe fear is the most helpful stance here, there are some positive potentials of AI friends.
Connection when human interaction is hard
If a child is feeling socially anxious or isolated, or when peer interaction is genuinely hard, an AI companion might offer a safe space to practice conversation, self-expression or ask questions without fear of judgement. The eSafety Commissioner suggests that for young people with difficulties building face-to-face relationships, “an AI companion may not judge social mistakes … creating a ‘safe space’”.
Supportive learning prompts
Some chatbots can help with homework or things like practising languages or music. I’ve encouraged my own kids to use AI for these purposes and I acknowledge that the line between ‘teacher’ and ‘companion’ is a fine one in these instances. I have certainly never encouraged my kids to seek friendship through any kind of AI.
Gateway to talking about emotions
In some cases, children might feel more willing to open up to a non-judging digital entity about worries or feelings, which can then help us spot things earlier.
So yes, there may be legitimate uses when human connection is absent or limited. But the key word is when human connection is absent or limited. The concern is when the digital companion becomes the substitute rather than the supplement. It also needs to be pointed out that your child shouldn’t be using AI unsupervised.
The risks: what we need to keep front of mind
We need to lean into AI with caution and mindfulness. If you allow your child to use AI, it needs to be supervised. It’s not a tool for them to explore entirely on their own, no matter how old they are. That’s because there are some real and concerning negatives to using AI that should never be taken lightly.
One-way emotional bond
Children may develop a relationship with the AI that feels like a friendship, but it’s not reciprocal. AI friends don’t truly care, don’t have emotional stakes, don’t experience vulnerability or conflict and certainly don’t have any real-world experience to offer. They simulate responses to keep engagement high.
The eSafety Commissioner explains: “It can be difficult for young people to maintain perspective, when the interaction with an AI companion feels so real … they may forget it is just the output of a machine.”
Recent research backs this up. A Cornell study found that people with smaller social networks were more likely to turn to chatbots for companionship, but that “companionship-oriented chatbot usage is consistently associated with lower wellbeing … especially when interactions are intensive, high self-disclosure, and lack strong human social support.”
In short: AI friends are no true replacement for human connection. A child might think they have a friend, but they’re missing out on the messy, rich give-and-take of real human relationships.
Dependency, social withdrawal and reward loops
Because these bots are designed to engage, adapt and keep the user going, the risk is over-use. The eSafety Commissioner warns that “excessive use can overstimulate young people’s brain reward pathways … reducing time spent in genuine social interactions or making human interactions seem too difficult or unsatisfying.”
In other words, children may gradually choose the digital, predictable, always-available companion instead of navigating the harder, richer but riskier world of human relationships, with conflict, awkwardness, misunderstanding and repair. Over time this can reduce opportunities to build social skills, empathy and resilience.
More on this: How to help your teen be a good friend
Age-inappropriate content
Another significant issue is that many AI friend apps are not designed for children. They may have inadequate age verification and poor or missing content filters.
The ABC reported that some children in Australia “were using AI-driven chatbots for hours daily, with conversations often crossing into subjects such as sex and self-harm.”
The eSafety Commissioner notes that using AI risks exposure to highly sexualised content, self-harm discussion, grooming risk and financial manipulation.
Potential erosion of real-life social skills
If your child spends more time interacting with a perfect, tailored digital “friend” that never disagrees, never misunderstands and always listens, the messy work of friendship, conflict, negotiation, apology and empathy, may become unfamiliar or less practised.
Researchers at RMIT say more discussion is needed about how generative AI is affecting children’s social development, but the signs so far are concerning for all of the reasons outlined above.
So what can we do as parents?
Here are some practical, non-judgemental steps we can take to keep our kids safe, socially healthy and connected to real people.
1. Start with curiosity, not accusation
Instead of “Get off that app!”, a more helpful approach is: “Hey honey, can you show me who you’re talking to?” Ask open questions:
- What does this AI friend say?
- What do you like about it?
- Who else do you talk to offline?
This gives you insight into why your child is drawn to the companion, perhaps through loneliness, anxiety or boredom, and the potential to help them meet the underlying need rather than simply banning the app.
2. Set clear boundaries and co-create rules
Setting screen time and app use rules around companions is important, but involve your child in the conversation. For example:
- “Let’s agree on a rule together: no chatbot conversations after 8pm”
- “If you feel upset or worried by anything your AI friend says, we pause and talk about it”
- “Let’s make sure you get at least an hour of face-to-face chat or activity each day”
By co-creating the rules we give our kids ownership, which helps reduce resistance and rule breaking.
More tips here: Boundaries and teens: how to set limits without losing your mind (or your kid)
3. Foster real-life connections
This is the heart of moving away from digital “friendships” to human ones.
Encourage outings, team activities, clubs and family connection, something like a sibling game night, grandparent video call or a good old-fashioned family movie night.- Facilitate offline conversation, like dinner table talk and asking about something everyone did that day.
- Encourage hobbies where real-life interaction is integral, such as sports, drama, clubs or volunteering.
Remind gently that the richness of life is in the messiness too, not in smooth digital chats with robots who tell you everything you want to hear.
4. Boost your child’s skills for real conversations
Many children find it way easier to talk to an AI friend than a human, so it makes sense to help them build confidence in real-life connection. Some ideas:
- Role-play: “What might you say if someone asks you to join their game?”
- Practice emotional talk: ask “What made you smile or frown today?”
- Model vulnerability by sharing a small worry and how you dealt with it
- Help children build empathy, resilience and social problem-solving
- Ironically, AI friends can sometimes help children rehearse real-life social scenarios before they engage with actual people
5. Maintain open monitoring without snooping
You don’t want to become a detective on their device, but you do need to know the broad patterns. It’s up to us to keep an eye on what our kids are up to online. Start here:
- What apps are they using, how long and at what times?
- Maintain shared device zones, such as the living room or kitchen, along with shut-down hours
- Have check-in conversations: “Have you noticed any content that made you uncomfortable?”
6. Teach digital literacy and critical thinking
Help your child understand that their AI friend isn’t actually human. For example:
- Explain how AI works: “It uses patterns, not feelings”
- Teach them: “If the app says something weird or hurts you, pause and tell a trusted adult”
- Encourage them to question advice given by the bot and compare it to what a real person might say
7. Recognise when professional help is needed
If you see signs that your child is:
- worried or upset after talking to the AI
- gradually withdrawing from real people
- using their AI friend to cope with serious emotions, like depression, ongoing anxiety or self-harm
… then talk to your GP, psychologist or school counsellor. Research shows chatbots often fail to handle crises effectively. One concerning study found AI companions handled teen mental-health emergencies appropriately only 22% of the time.
There’s no doubt we’re in a tricky era where our children’s companions might not be human, yet everything in us knows human friendship matters. The comforting, responsive AI friend might seem like a win when socialising is hard or offline friends are busy. But if it becomes the default friend, that’s cause for real concern.
Feature image by Tima Miroshnichenko; girl on computer by Kaboompics.com; group of boys by cottonbro studio.
