Nearly half of Canadians turn to AI for medical advice: Canadian Medical Association

Would you trust artificial intelligence (AI) for medical advice and treatment?

Nearly half of Canadians now say yes, according to a new study by the Canadian Medical Association (CMA) raising concerns among doctors.

Most Edmontonians told CityNews they were  skeptical of AI when it came to medical advice as well.

Sierra Wallace said, “AI is wrong often. It makes mistakes. I wouldn’t trust it with my personal information.”

Another Edmontonian Avijot Brar said, “They don’t really know as much as a doctor. They don’t have that experience, and they could pull information from fake studies.”

However, some like Britney Mackinnon see some benefits to using AI as tool for research and basic treatment.

“People might find that controversial, but like I mean, if you’re using it the same way as like you’re using, say, a search engine, trying to determine whether or not it’s a good idea to take Tylenol and Advil together, it’s probably fine,” Mackinnon said. “The answer to that is it’s okay, actually.”

In the survey by Abacus Data, 48 per cent of Canadians said they use AI like ChatGPT for health information, while Google was at 52 percent.

When it came to looking for treatment, 39 percent rely on ChatGPT and 48 percent in Google.

The president of the medical association encourages people to connect with professionals like family doctors or Alberta 811.

“They know your full story. They know your history, your medications, and they can work with you to establish the correct diagnosis and treatment,” said Margot Burnell, president of CMA.

But some Edmontonians say the lack of doctors across the province is leading people to rely on AI — or the internet — for medical advice due to speed and convenience. 

“That’s kind of scary and unfortunate,” Wallace said. “My guess is that it’s probably in part due to the lack of family doctors.”

“Nobody really knows what information is good and what information is bad, and nobody knows how to parse through all of it, so it makes sense that turning to a single model or a single tool might be able to give you some semblance of advice it totally makes sense,” Mackinnon said.

Burnell admitted the technology will improve and it can be useful in the future, but false information poses a risk for people seeking help, she said.