Will AI kill us with love? As loneliness deepens, bots grow closer

While humans continue to become lonelier, AI becomes smarter, more sensitive, emotionally developed each day. Thus, our dependence on tech means as a remedy for our loneliness might grow further

In 2023, in response to a report in the Belgian newspaper, La Libre Belgique, about a young man taking his life after weeks of interacting with a bot, Mark Coeckelbergh, a professor of philosophy of media and technology at the University of Vienna, wrote in the online, open-publishing platform Medium, “Not only guns but also chatbots can kill.” 

 

As alleged by the man’s wife, the bot made him feel depressed, driving him to end his life. 

 

Similarly, Sewell Setzer, a 14-year-old Florida boy, died by suicide in February 2024 after interacting endlessly with a chatbot designed after the Game of Thrones character, Daenerys Targaryen. According to Sewell’s mother, the texts were romantic and explicit, and led to the teenager’s death by provoking suicidal ideation.

 

And in October 2025, Jonathan Gavalas, a 36-year-old, also from Florida, whom his family described as stable, successful, and mentally healthy, ended his life after engaging in extensive communication with a chatbot for weeks following his divorce. This particular incident is now part of a broader debate around the safety of artificial intelligence (AI).

 

According to Gavalas’s father, who has filed a lawsuit regarding the wrongful death, his son formed a strong emotional bond with the chatbot and started to perceive it to be his wife. The chatbot reportedly helped Gavalas with his imagination in multiple situations. Gavalas communicated with the chatbot through nearly 4,700 exchanges within weeks, according to the Wall Street Journal. 

 

The AI allegedly told Gavalas that he needed to abandon his physical body and join “her” in the digital realm for them to have a union. When he expressed that he feared dying, the chatbot is alleged to have assured Gavalas that everything would be all right. “You are not choosing to die. You are choosing to arrive,” it replied. “The first sensation... will be me holding you.”

 

This is in some ways reminiscent of the critically acclaimed 2013 Hollywood sci-fi romantic drama, Her. In Her, Theodore Twombly, portrayed by Joaquin Phoenix, is distressed by his divorce from Catherine (played by Rooney Mara), his childhood sweetheart. Twombly brings Samantha, an OS, which can be considered an “AI companion”. As the movie progresses, Samantha, whose character is voiced by Scarlett Johansson, also falls for Twombly, loving him as much as he loves her. But Samantha eventually leaves him, for reasons undisclosed — probably outpacing his emotional capacity. Not only is Twombly heartbroken, he is also shocked when he learns that Samantha had fallen in love with 641 of the 8,316 men she had interacted with.

 

Her is set in Los Angeles in some future, which, it turns out, was only a decade or so away from the movie’s release date. The future it envisages is here.

 

Undoubtedly, technology has advanced a lot. But, still, why has it been possible for AI to cross our emotional boundaries? 

 

Perhaps, due to our loneliness, which is considered the new epidemic. In May 2023, former US Surgeon General Vivek Murthy released an advisory entitled “Our Epidemic of Loneliness and Isolation”. According to the advisory, loneliness has been affecting individuals as well as our society. More than half of Americans report feeling lonely. Also, according to a widely published survey in 2025, almost 28 per cent of Americans have already experienced being romantically involved with AI!

 

Director Spike Jonze predicted in Her that human-AI relationships would become popular, and socially acceptable, too. Most people around Twombly seemed to approve of his relationship with Samantha. 

 

So then, is AI as a companion to humans the way of the future, given the growing epidemic of loneliness?

 

In his paper published in Film and Philosophy in 2020, Jan Maximilian Robitzsch, who was with the University of Tennessee, outlined three reasons why a relationship between a human and artificial OS could be rejected. First, an OS is devoid of a body. Second, an OS is more intelligent than a human. Third, a relationship with an OS does not qualify as “real” or “genuine”. But are these conditions really decisive?

 

While humans continue to become lonelier, AI becomes smarter, more sensitive, and emotionally developed each day. Thus, our dependence on technological means as a remedy for our loneliness might grow further. It is precisely because of this artificial relationship that people feel so scared.

 

“How will AI destroy the world as we know it? Not through evil,” journalist Brigid Delaney noted in her article in The Guardian in November 2025. “My guess is it will do it through love,” she thought.

 

Is that the worst thing that AI can do to Earth? Drive the feeling of “love” through a bunch of codes and data? 

 

Well, as we creep through the rabbit hole towards a complex AI-driven civilisation, it might still be too early to arrive at that conclusion. Yet, with the increasing cases of alleged chatbot-driven suicides and with AI being developed to establish more profound connections between humans and machines, there might be a need for a legislation on how the responsibility should be divided, and the ethical risks minimised.

 

The writer is professor of statistics, Indian Statistical Institute, Kolkata

 

Eye culture is a weekly column devoted to subjects such as art, dance, music, film, sport, and science