Users are reporting bizarre replies from AI chatbots. We want to hear yours

A man looks at the new iPhone 16 in the Apple store as the iPhone 16 with artificial intelligence software, and other Apple Watch products, go on sale, in Berlin, Friday, Sept. 20, 2024. (Katharina Kausche/dpa via AP)

Modern, human-like AI chatbots have arrived, and some mental health experts are raising the alarm.

With the advent of new AI tools powered by large language models (LLMs), the market for chatbots designed for conversation, companionship and even something similar to therapy have shown substantial growth.

But turning to artificial intelligence in times of loneliness or poor mental health may pose risks.

Researchers from Stanford University recently found what they call “foundational and practical barriers to the adoption of LLMs as therapists,” concluding in an April conference paper that AI tools should not replace traditional, human counsellors.

And in a recent study analyzing thousands of user reviews for a companion chatbot, researchers from Philadelphia’s Drexel University uncovered reported stories of crossed boundaries and misconduct, including what researchers called “AI-induced harassment.”

Have you had a conversation with a chatbot that you think has affected your mental health? Have you considered turning to artificial intelligence for companionship or counselling? Are you concerned about the emotional impacts of AI on someone you know?

We want to hear from you.

Share your story by emailing us at [e-mail] with your name, general location and phone number in case we want to follow up. Your comments may be used in a CTVNews.ca story.

With files from CTV News’ Tammy Ibrahimpoor, and from The Associated Press