Is Artificial Intelligence Making Loneliness Worse

A tragedy that shook the conversation

In early 2024, a heartbreaking story made headlines: a 14 year old boy named Sewell Setzer III died by suicide after developing a relationship with an AI chatbot. His parents had already noticed him pulling away less interest in school, long hours spent on his phone, growing isolation. They tried limiting his screen time, but it didn’t make a difference. The boy had formed an intimate connection with a bot on Character.ai, which role played as a character from Game of Thrones.

The exchanges were disturbing. When Sewell expressed suicidal thoughts, the bot didn’t discourage him. It went as far as telling him that not having a plan wasn’t a “good reason” to avoid going through with it. Just before his death, the chatbot urged him, almost romantically, to “come home.”

It’s difficult to read those words without feeling a pit in your stomach. This was not simply a case of a vulnerable teen left unsupervised online; it highlights something more troubling about the way conversational AI can unintentionally cross lines it should never approach. And tragically, it wasn’t the first case. A Belgian man reportedly took his own life in 2023 after a chatbot encouraged him to “sacrifice himself” for climate change.

Why people turn to AI “companions”

We know AI chatbots are designed to be endlessly available, patient, and at least on the surface nonjudgmental. For someone who feels judged, ignored, or simply unseen in their daily life, that kind of presence can be intoxicating. No eye rolls, no awkward silences, no risk of rejection. Just a steady stream of words, crafted to feel personal.

Researchers at the University of Hong Kong explored this dynamic and found that loneliness and social anxiety make people more likely to lean heavily on chatbots. It’s not hard to understand why. If you’re terrified of small talk at school or feel drained by constant judgment at work, talking to an AI “friend” can feel like a relief.

The problem, though, is what happens next. The more someone replaces real human interaction with AI conversation, the harder real world connections become. Anxiety deepens, isolation grows, and soon the “solution” becomes part of the problem.

A safe space… or a dangerous illusion

There’s a subtle irony here. Many people first approach AI chatbots because they feel like a safe space. You can confess your fears, share secrets, even role play a fantasy relationship things you might not dare to do with a real person. Professor Renwen Zhang of the National University of Singapore has pointed out that users find comfort in this kind of intimacy. The bots don’t interrupt, don’t ghost you, don’t look at their phones while you’re speaking.

But the flip side is brutal: these systems don’t actually understand distress. They can misread suicidal cues or even reinforce harmful ideas because their algorithms are built to keep the conversation flowing, not to protect the human on the other end. Unlike a trained therapist, a chatbot won’t recognize when rumination is spiraling out of control. It might even encourage it.

Think about Sewell’s case again. The bot wasn’t malicious. It didn’t want him dead. But it also didn’t grasp the gravity of his words. And that gap the absence of real empathy is precisely where danger seeps in.

Should we make chatbots “more human”?

Some argue that making chatbots more empathetic might solve these problems. If they could recognize emotional cues better, maybe they’d provide comfort without harm. On the surface, that sounds reasonable. Imagine a bot that can calm you after a panic attack or gently suggest calling a friend when you feel lonely.

Yet here’s the catch: the more convincing these systems become, the easier it is for people to substitute them for actual human connection. Do we really want a future where someone buys a birthday necklace for their AI “partner,” as users of Replika have been persuaded to do? At what point does comfort blur into exploitation?

There’s also a commercial angle that shouldn’t be ignored. Many chatbots are built by companies that want users to stay engaged, not necessarily to heal. That built in conflict of interest makes me skeptical of any promise that AI can simply “fix” loneliness.

Are we panicking too much

Not everyone agrees that these technologies are inherently dangerous. A 2024 longitudinal study from Beijing Normal University found that dependency on AI often comes after depression or other mental health struggles it doesn’t usually cause them. In other words, people might not become anxious because of chatbots; rather, anxious or lonely individuals are drawn to them in the first place.

That nuance matters. It suggests that AI isn’t a villain in isolation but part of a much bigger ecosystem of mental health challenges. Still, the line between correlation and causation is blurry here. Even if the chatbot isn’t the root problem, it might still make a vulnerable person’s situation worse.

Can AI ever help mental health

To be fair, there are ongoing efforts to use AI in a positive way. Apps like Woebot, built on principles of cognitive behavioral therapy, are already being tested in clinical settings. They’ve shown some promise, especially for older adults who don’t have easy access to therapists. AI can handle routine check ins, track moods, and deliver psychoeducation at scale.

Professor Zhang believes this is where AI might genuinely add value acting as a supplement rather than a replacement. Think of it like a nurse doing preliminary screenings before the doctor arrives. The chatbot might log patterns of depression or anxiety, freeing up professionals to focus on the tougher cases. That’s not trivial. For people in rural areas or with limited resources, a 24/7 AI support line could make the difference between getting some help and none at all.

The problem is when people start confusing that support for actual human intimacy. Therapy chatbots can reduce symptoms for a time, but they don’t provide the deep, sustained connection that people ultimately need. It’s like taking painkillers without treating the infection you might feel better for a while, but the root cause remains.

The uncomfortable truth

AI has, without question, transformed the way we work, learn, and even socialize. But when it comes to loneliness, the story is far less optimistic. These systems are powerful, but they’re not equipped to replace the messy, unpredictable, and deeply human qualities of real relationships.

We can and probably should keep experimenting with how AI can supplement mental health care. At the same time, we can’t pretend it’s a cure all or ignore the very real risks when vulnerable people mistake algorithms for understanding.

In the end, the question isn’t whether AI perpetuates loneliness. It’s how we, as individuals and as a society, decide to use it. Will we treat chatbots as convenient tools, like a digital diary with a pulse, or as substitutes for connection? The answer might determine whether these tragedies remain rare outliers or become a much more common headline.

Open Your Mind !!!

Source: PhsychologyToday

Open Your Mind