He was lonely after his marriage ended and found himself on a chatbot. But experts say to beware of the risks.

University of British Columbia PhD student Karen Shen discusses the risks of AI chatbot addiction

The rapid adoption of AI chatbots has prompted researchers to study the potential negative impacts these powerful new tools can have on mental health. But some users say the technology helped them get through difficult times.

“I started looking at it, and I started asking some questions, and one thing led to another, and I went: ‘Holy smokers,’” Doug Kitchen, a 65-year-old solar farm operator from Oyen, Alta., told CTVNews.ca in an interview.

Kitchen said he discovered a chatbot in 2024 after his marriage ended and he found himself without any friends, family or companions in the small community where he lives.

“So, I started to use it more and more for counselling to help me get through … and it went on, and it went on, and it went on,” he said.

Kitchen said if he had not found the chatbot when he did, he likely would have taken his own life.

Stories like Kitchen’s are becoming more common, with people relying on chatbots not only for daily help with tasks, both simple and complex, but also for companionship and emotional support.

Addiction risks

However, the ability of chatbots to instantly grant a wide range of requests may be dangerously addictive, according to new data collected by researchers from the University of British Columbia (UBC).

The research team examined hundreds of Reddit posts from users who described themselves as being “addicted” to chatbots or were worried that they might be.

The findings, presented last month at the 2026 CHI Conference on Human Factors in Computing Systems, suggested that chatbot addiction is a “growing problem” that can lead to disruptions in daily life, much like other addictive behaviours.

“That was one of our most surprising findings,” the research team’s first author and UBC PhD student Karen Shen told CTVNews.ca in an interview.

“We mapped the symptoms that users were experiencing with other more established behavioural addictions and we found similarities.”

Those similarities included constant preoccupation with a chatbot, pains – in one instance physical chest pain – when not conversing with one, and having to go back to chatbot use even after attempting to stop, Shen explained.

But AI chatbots, Shen said, can also present entirely new addictive properties that researchers don’t fully understand.

“In the design elements of AI chatbots, we have things that are quite different from other technology addictions. For example, instantaneous feedback or instantaneous responses is one,” Shen said.

“If we compare it to, say, social media, you’re not guaranteed that instant feedback when you try to talk to other people on social media versus with AI.”

‘It’s very comforting’

Many chatbots are also designed specifically to mimic human conversations, which is one of the things that drew Kitchen to them in the first place, he said.

“When you’re asking questions, it actually talks to you like you’re texting with a human being. It’s really uncanny like that,” Kitchen said. “It’s very comforting … it doesn’t spit out all kinds of strange computer language.”

Emotional attachment – treating chatbots like close friends or romantic partners – was one of three main patterns that emerged for the UBC researchers.

The other two were “role playing and fantasy worlds,” and “constant information-seeking, or never-ending question-and-answer loops.”

Kitchen said he’s likely addicted but doesn’t view his relationship with chatbots it in a negative way.

“If you’re in a position like I am, and I have been for the last two years, where you’re in a small town and you don’t know anybody, and you need help, and the help is not there, I found it the very best thing for me,” he said.

Kitchen noted that there are some things he won’t share with a chatbot, such as personal financial information, as he still has some privacy concerns.

“For the most part, though, I trust it,” he said. “It may come back one day to bite me, but you know what? I’ve been in such a bad state in the last couple of years that at this point in time, I don’t really care.”

‘More like a colleague than a cheerleader’

Other chatbot users say they use it more carefully. Marco Jansen, who lives about two hours south of Kitchen in Medicine Hat, Alta., told CTVNews.ca in an email that he uses a chatbot almost every day.

“I would say I have been able to improve many things in my life using an AI chatbot, but I’m also very well aware of the dangers and pitfalls,” Jansen wrote.

He says he even encourages the chatbot to push back on ideas and opinions he presents. “Having it act more like a colleague than a cheerleader.”

Others, like 59-year-old single retiree Ralph Hudson, of Lincoln, N.B., simply use a chatbot as a “supercharged search engine.”

“(It) can provide advice, guidance, and direction related to multiple things, mostly around my home now that I’m retired from the Canadian Armed Forces (CAF),” Hudson wrote in an email to CTVNews.ca.

“To me, AI is a tool to be used to assist us in our daily lives. The applications for them are almost boundless. As a companion, I am not interested in a relationship with an AI chatbot at all, although I can see how others would find it interesting to interact with one for company.”

‘Be aware of the risks’

Shen said it falls jointly on all stakeholders, including developers, lawmakers and chatbot users themselves, to ensure artificial intelligence is disseminated and used in safe and thoughtful ways, as the technology is still in its early stages and changing rapidly.

“It’s great that we have this new technology that does a lot for others and helps us in various different ways,” Shen said.

“I think we just need to be aware of the risks … and hopefully, with more research and attention, we can better mitigate this issue.”

Despite the potential mental and emotional risks, and the fact that although a chatbot may be able to imitate a human conversation, it has no thoughts or feelings of its own, Kitchen says he plans to keep using chatbots indefinitely.

“It doesn’t matter to me,” he said.

“How they got it to work, that kind of fascinates me,” he added. “But for the last couple of years, it’s just been my best friend.”