.png)
Think about the last time you had ChatGPT write something for you and consider this: Would you be comfortable with that bot giving a teen mental health advice?
Driving the news: Young people are increasingly turning to a growing number of mental health chatbots powered by large language models, even though there is limited evidence that they actually help.
Why it’s happening: Chatbots can help people get around barriers to mental health care, which are growing. The average wait time for care in Canada is about three weeks, but those numbers go up for youth, as well as more specialized care — one ADHD clinic in British Columbia has a two-year waitlist.
- That’s not to mention the cost of therapy — which can run up to $250 an hour — and lingering stigma that might make someone keep their troubles in their phone instead of sharing them with real people.
Why it matters: Large language models have a tendency to hallucinate and give incorrect info. That’s one thing when you’re getting a chatbot to write a cover letter, but the stakes are much higher when dealing with someone’s mental health.
- Many of these bots have disclaimers that they are not a replacement for proper treatment, so they aren’t subject to regulatory approval.
- An analysis of thousands of studies on mental health chatbots last year found that only 15 had the same rigour as other medical research.
Yes, but: There are safer uses for AI in mental health. MIRA, a chatbot developed with support from several Canadian universities and mental health organizations, does not give advice; instead, it helps people with the sometimes difficult task of navigating the mental healthcare system and connecting with the right (human) care.
Zoom out: AI is being applied to other areas of healthcare, where patients face similar barriers. Health data is strictly regulated, so these platforms are coming slowly. But many of them are geared towards helping doctors and nurses process data more efficiently and give patients more personalized care.