In 2000, Robert D. Putnam published Bowling Alone, his landmark account of America’s retreat from communal life — from churches and civic groups to social clubs and after-work bowling leagues. The consequence, he argued, was a steady erosion of “social capital”: the networks, norms and habits of trust that bind individuals together.

Were Putnam to republish the book in 2026, it might warrant a darker sequel: Talking Alone. We have not merely withdrawn from public social rituals but from private ones as well. Social media, far from remedying isolation, has become profoundly antisocial. Increasingly, it trains us to passively consume AI-generated content rather than sustain relationships with other people. Why invest time and emotional effort in fallible, flesh-and-blood friends when a machine offers unlimited attention, patience, and empathy?

AI’s invasion of interpersonal relationships is now influencing children, and a new report has found that 81% of those aged 11-16 in the UK use AI chatbots. Additionally, 40% do so because they are lonely, while 31% describe AI as similar to a friend, and 24% say that they would use it for advice on difficult situations. One in three has shared information with chatbots that they would not tell their teachers, parents or friends.

Teenagers divulging their innermost thoughts to an algorithm for an average of 40 minutes a day is not the same as them confiding in an imaginary friend. These companion services are cynically exploiting loneliness for commercial gain. They are about profit and maximum user engagement, and so will train themselves on our desires to be as seductive as possible.

AI is built to be sycophantic, with its craven desperation to unconditionally agree with everything a user says. In turn, this creates a fawning feedback loop in which people are protected from any risk of judgement or emotional conflict. It not only “deskills” young people, who may lose the ability — or motivation — to socialise in real life, but produces a one-way relationship which indulges our most introspective, navel-gazing impulses.

Technology allows us to be even more solipsistic than we are already. In 2024, actress Jemima Kirke did a Q&A session on her Instagram stories. When one of her followers asked her what advice she’d give to unconfident young women, she replied: “I think you guys might be thinking about yourselves too much.” It’s become something of a meme, encapsulating how modern-day self-awareness has actually just become about self-involvement.

The fear remains that AI chatbots will exacerbate our main-character-syndrome obsession. Having an automatic, appeasing response to every insecurity spiral, every opportunity for overthinking, every moment of downtime or boredom is not good for anyone’s mental health.

Yet this is alluring — and therefore dangerous — for children and teenagers, who are navigating “big” emotions and experiences without the life experience which helps them differentiate fact and fiction. What they need from AI is something they will never receive, because it goes against the entire business model: pushback. They need reminding to look up, to talk to someone human, to go outside, and just to turn the damn thing off.

---

Kristina Murkett is a freelance writer and English teacher.

kristinamurkett

---