Could AI companions change how young people relate to each other in the future? Liz Bury takes a look.
These days, a growing number of children are coming home from school, picking up their phones and logging onto devices to talk to AI chatbots. Recent research by Internet Matters found that two-thirds of UK children aged 9 to 17 use them regularly, and critically, a third of those users consider the chatbot to be a friend.
What happens when young people are raised on relationships where there’s no friction, their ideas are never questioned or there aren’t any non-verbal signs they need to interpret?
As digital companions become increasingly common in childhood, they may already be reshaping what the next generation unrealistically expects from real-world human relationships later in life.
False friends
Much of the recent conversation around online harms has focused on the inappropriate material children might see. But the AI companions issue takes the level of concern up a notch. These personas are built to be always pleasant, attentive and accommodating, as well as being available to their users 24/7 – qualities that no human peer friendship can match.
Real relationships with people will always contain friction and disagreements on some level, and part of being human is learning to navigate those tensions during our formative years. Chatbots are simply not designed this way; after all, the developers want to keep people coming back for more. Professor Hannah Fry’s research highlights a troubling pattern now emerging, which is that people who rely heavily on AI companions report feeling “understood” by them but subsequently find real-world friendships frustratingly unpredictable. When an AI app always flatters and validates you, that relationship is going to feel very easy in comparison.
This type of constant positive reinforcement has the potential to create a feedback loop where children won’t know what to do when someone airs an opinion that’s different to theirs. To get a clearer picture of where we could be headed, Professor Fry’s documentary explores exactly how these dynamics unfold in practice and why we should take note now. It’s a compelling watch and I recommend it if you haven’t seen it.
Regulation is behind the curve (again)
At this point, standalone companion apps such as Character.AI, Replika, and Candy AI have tens of millions of users between them. It’s already past the point of being able to regulate them responsibly before they landed in the hands of children.
The UK government recently announced that, as part of measures affecting under-16s’ social media access, AI “romantic companion” chatbots must enforce a minimum age of 18, with intimate functions restricted for under-18s. All good as far as it goes, but there are two gaps. Firstly, the onus is on the developers to self-enforce these limits – and we haven’t seen this go well on other platforms so far. Secondly, this only covers the romantic companion bots. What about general emotional support, medical or conversational AI tools that build similar relationships and dependencies?
The question is whether regulation will address the deeper problem of interference with social learning. A ban on romantic chatbots may limit exposure to under 18s, but unless we also help young people understand why they’re drawn to these relationships in the first place, we risk pushing them elsewhere to find them and not solving the problem at all.
Staying vigilant
There’s no single red flag that signals a student is becoming dependent on AI chatbots. It may be something as simple as increased talk about an AI friend or withdrawing from social interaction with classmates. Schools can remind staff to be alert to these changes, alongside maintaining the safe spaces where young people feel they can discuss their online relationships without being judged.
Social skills aren’t optional and, unfortunately, the burden is increasingly falling on schools to help where technology is sabotaging the basic skills we should all be learning intuitively.
Students must understand that human relationships can be complicated yet amazingly rewarding – but they need to put the work in to benefit and it’s non-negotiable if they are to live happily in the real world.
It takes effort in the face of tempting distractions like AI ‘friends’ where it’s all just too easy.
Looking ahead
Once again, the imperfect technology is not going away. And the onus is on schools and teachers (also, once again) to equip their students to understand the full picture in a fast-moving tech environment. Staying vigilant in the real world, educating students and having honest conversations about human relationships are all ways to ensure they have positive and fulfilling friendships and relationships with others in the future.
References
- Internet Matters, “Me, Myself, & AI” report. Available at: https://www.internetmatters.org/hub/press-release/new-report-reveals-how-risky-and-unchecked-ai-chatbots-are-the-new-go-to-for-millions-of-children/ (Accessed 25 June 2026)
- AI Confidential, with Professor Hannah Fry. Available at: https://www.bbc.co.uk/iplayer/episode/m002q76d/ai-confidential-with-hannah-fry-series-1-1-the-boy-who-tried-to-kill-the-queen (Accessed 25 June 2026)
- Social media to be banned for under-16s in landmark government move to give kids their childhood back. Available at: https://www.gov.uk/government/news/social-media-to-be-banned-for-under-16s-in-landmark-government-move-to-givekids-their-childhood-back (Accessed 25 June 2026)
