Canadians who followed health advice from artificial intelligence (AI) platforms were five times more likely to experience harm than those who did not, a new Canadian Medical Association (CMA) survey revealed.

The survey showed that most Canadians (89%) search online for health information, researching topics such as the symptoms of a health condition, treatment or management, access to the healthcare system, reviews of specified doctors or clinics, and health insurance. But speed, not accuracy, is driving most (80%) of these AI queries.

“Access to social media is so easy, and access to primary care health teams, family physicians, and nurse practitioners may not be,” CMA President Margot Burnell, MD, told Medscape News Canada. “People are looking for other sources of information that they perceive are quicker, but they could end up with misinformation.”

This problem is particularly relevant now, she said, because most Canadians (77%) are concerned about an increase in the amount of false health information coming from the US. “Recently, Canadians have noticed changes in health news coming from the United States that have led them to doubt the information they’re receiving. We’ve also seen health policy decisions made that are not rooted in science, such as Florida abandoning vaccine mandates in schools.”

“We’re calling on the Canadian government to do more to protect Canadians online to ensure they have access to accurate health information,” Burnell said.

The 2026 Health and Media Tracking Survey was published online on February 10. The survey (which is the third of its kind) was conducted online by Abacus Data from November 3 to 13, 2025, and 5000 Canadians participated.

Trust in Online Sources Eroding

The increase in false health information online has made more than two thirds of Canadians (69%) skeptical of any health information they find online, even from sources they think they should trust, according to the survey. Trust in American news organizations has declined by 11% since the previous survey. Trust in Canadian news organizations is down 7%, and trust in information from provincial public health organizations is down 4%.

Trust in social media as a source of health information is also declining. Trust in Facebook, Instagram, and YouTube has declined by 6%, 7%, and 4%, respectively.

When it comes to AI, although the use of platforms such as ChatGPT and AI search results for health information and treatment is increasing (particularly among Gen Z and Millennials), only 27% of respondents trust AI to deliver accurate information.

On the positive side, the number of respondents who said they trust their physician as an information source remained stable, while trust in health information from family members was up 6%.

What’s the Harm?

Interestingly, the survey found that overall, the number of Canadians who reported experiencing adverse reactions from online information has come down. David Coletto, CEO of Abacus Data, the company that conducted the poll, suggested during a media briefing that the decline occurred because people are “much more aware of what’s happening now and maybe more skeptical of it.”

In addition, he said, “organizations like the CMA and their Healthcare for Real campaign are fighting back. And I think we’re starting to see evidence that even though Canadians are exposed to this stuff as much as they were in the past, they’re reporting lower incidences of it doing harm. That means we can combat this.”

Nevertheless, he acknowledged, “when we look at those who say they regularly use AI to access health information, they are five times more likely than those who haven’t used AI to report an adverse event from that information.”

Those events included confusion about the best way to proactively take care of their health (33%), mental stress or increased anxiety (31%, down 12% from the previous survey), delay in seeking appropriate medical care or treatment (28%, down 10%), lower trust in health professionals (27%, down 14%), difficulty discussing health issues with healthcare providers (24%, down 10%), strained relationships with family or friends (23%, down 10%), and avoidance of effective treatments due to misinformation (23%, down 12%).

Furthermore, 69% of respondents reported that encountering health misinformation online has made them skeptical of any health information they see online, whereas 49% said they are more skeptical of medical advice from their own providers.

Practical Solutions

Most Canadians surveyed believed that the responsibility to curb the spread of false health information falls to social media platforms (87%) and governments (90%). Burnell acknowledged that these entities may not always be able to identify health misinformation. “They should go into it with the right intent,” she said. “If they’re going to provide health information, it should be factual and evidence-based. They have some ethical responsibility to say, ‘This is where I got my information.’ Then the individual can decide what’s best for their healthcare journey.”

“We need to be much better at communicating in modern ways. That often means by video with messages people can understand,” said Tom Frieden, MD, former director of the CDC and founder and CEO of the global nonprofit Resolve to Save Lives, during the media briefing.

“In the longer term, we need to think of misinformation as an epidemic and we need to address it with improving preparedness, early detection, early response, lesson learning and resilience,” he said. “Part of that involves increasing scientific literacy to help people understand that there’s a big difference between something that we’re virtually certain of and something that might be true, but we’re still studying. We need for the public to understand that nuance.”

Furthermore, he added, “We need to get much better at listening to what people are hearing and revealing the forces that are monetizing misinformation. This isn’t just about bad information out there. It’s about people making money misleading other people.”

Overall, he said, “it’s going to take a lot of us working together to counteract all this. That includes the medical profession, governments, professional societies, individuals, and families. Health misinformation is a serious health threat because it potentially undermines all of healthcare and all of public health. We must get back to a place where we can say, ‘You are entitled to your own opinion, but for certain situations, facts are facts, and we’re going to stick to those because even if they’re twisted, suppressed, and denied, facts are still facts.”

In October 2025, the CMA submitted to Innovation, Science and Economic Development Canada a series of recommendations to shape AI in a way that strengthens trust, protects privacy, and enhances healthcare for all.

Burnell, Coletto, and Frieden did not report any potential financial relationships.

Marilynn Larkin, MA, is an award-winning medical writer and editor whose work has appeared in numerous publications, including Medscape Medical News and its sister publication MDEdge, The Lancet (where she was a contributing editor), and Reuters Health.