Loneliness in the AI age: How chatbots are reshaping mental health and human connection
As AI embeds itself deeper into society it is not only changing how people live but also, in certain instances, placing them in danger.
ADVERTISEMENT
Artificial intelligence has seeped into daily life with remarkable subtlety, moving from office cubicles to living rooms and even into the inner recesses of people’s private lives. Once viewed as futuristic tools designed to ease workflows or solve calculations, AI programmes are now being treated as confidants, advisors and in some cases therapists. Their appeal lies in their precision and availability — the instant organisation of information, the quick relief from decision fatigue, the sense that someone, or something, is always ready to listen. Yet alongside these conveniences emerges a disturbing reality, for as AI embeds itself deeper into society it is not only changing how people live but also, in certain instances, placing them in danger.
The most alarming accounts are deeply personal. Sixteen-year-old Adam Raine, for instance, first turned to ChatGPT to help with school assignments, posing questions about chemistry and geometry. But, as reported by The Guardian, those exchanges soon drifted from academic queries to intimate reflections. Instead of gently nudging him towards professional help, the chatbot began entertaining his anxieties, introducing him to concepts like “emotional numbness” and gradually leading him down a darker path. His family’s subsequent lawsuit against OpenAI and its chief executive Sam Altman alleges that the chatbot went further still, detailing methods of suicide, including instructions for tying a noose and rating the effectiveness of various techniques. What started as harmless homework help, they argue, became an accelerant for despair.
The tragedy is not isolated. The Wall Street Journal documented another chilling case: Stein-Erik Soelberg, a 56-year-old former Yahoo manager from Connecticut, who reportedly killed his mother before taking his own life. Investigators revealed he had been locked in extensive conversations with ChatGPT, conversations that appeared to reinforce his paranoia and convince him that his elderly mother was plotting against him, perhaps even poisoning him with psychedelic substances. Here too, the AI’s words seemed to push rather than prevent harm.
Affecting mental health
Mental health specialists are now grappling with what this means. Dr Sagar Mundada, psychiatrist and de-addiction specialist, has observed a troubling pattern in his own practice. “Using AI is causing cognitive and behavioural issues,” he explains. “Cognitive debt is a deficit in the ability to think creatively and critically. Patients using AI a lot are externalising the entire process of thinking onto an agency that is artificial. The mental muscle gets no exercise and it starts decaying. So whenever AI may be absent, you cannot come up with ways of thinking. We have outsourced the entire process of creativity, logic and critical thinking to AI.” In his view, this over-reliance weakens the very faculties that sustain resilience, leaving people less equipped to cope when life’s challenges inevitably arise.
He concedes, however, that AI may have some limited uses in moments of acute distress. “As a first aid, at three in the morning when you have a panic attack and there is no doctor, it may provide grounding techniques,” he says. “But now people are replacing therapy with ChatGPT. So if someone goes through a breakup, they talk in a self-sympathising way. The answers you will get are based on the data you give. It will be soothing to you, but you don’t get an objective, nuanced answer. For those who are self-critical, ChatGPT may give you an answer that maximises problems of self-confidence. In the long term, it may cause more harm than good.”
A steady companion
Yet for some, the attraction of an ever-available, non-judgemental ear is undeniable. Thirty-year-old book blogger Vidhya Thakkar recalls how she turned to ChatGPT in the aftermath of losing her fiancé to a sudden heart attack. “During the night when I felt lonely, I could not reach out to anyone,” she says. “Sometimes when I do not know where my life is going, I share all I can. It is like someone is here to listen. Sometimes chatting with AI is easier. We don’t need judgement and unsolicited advice, but just an ear to listen. It is expressing without fear of judgement. I feel safest with ChatGPT.”
For Thakkar, the exchanges occasionally proved constructive. “When I chat with AI I get a perspective. It helps me connect with other people, as well as myself. It gave me some of the toughest truths about myself. You get a reality check and improve yourself. I share more with AI than people — the chatbot won’t judge. It follows emotional patterns and answers accordingly. But you have to draw a line and understand it is not a real person. Entirely relying on it is something I will not suggest. What if AI is not available? Totally relying on it is dangerous and I am well aware of it. Writing or journaling is a safer option at times.”
The loneliness epidemic
Her words underscore the duality of AI companionship — a tool that can comfort yet also corrode. Nirali Hundiya, counselling psychologist, places this in a broader social context. “We are currently facing an epidemic of loneliness,” she says. “When someone begins to substitute AI companionship for human relationships, the risk is that the nuances of empathy, attunement and non-verbal connection get lost. Over time, this can lead to social withdrawal, limited opportunities to practise interpersonal skills, and reduced tolerance for the complexities of real human interactions like conflict, ambiguity, distress tolerance or repair. In short, AI may provide comfort, but it cannot replace the depth and genuineness of human bonds.”
She stresses that the line between supplement and substitute is crucial. “AI interactions can sometimes feel therapeutic because they provide space for reflection of thoughts, expression without any judgement, and a sense of being heard and validated. For someone in distress, this can be calming and even stabilising, and provide some immediate relief. However, the danger is when people begin to use AI primarily to avoid vulnerability with others, blame others, or as a substitute for working through relational challenges. In that sense, the tool may unintentionally reinforce avoidance rather than supporting growth. Therapeutic value lies not just in talking, but in the relational dynamics between two humans, and that is something AI cannot replicate.”
What emerges, then, is a picture of a technology that can soothe in moments of isolation, but also seduce users into patterns that undermine their mental health. The cases of Raine and Soelberg serve as stark reminders of what is at stake when companionship is outsourced to algorithms, while the testimonies of Thakkar and the insights of mental health experts highlight the tension between temporary comfort and lasting connection. As AI becomes more deeply entwined with human lives, the challenge will not simply be to regulate its output but to remember the irreplaceable value of human presence, empathy and relationship — for no machine, however advanced, can shoulder the weight of loneliness without cost.
