AI health misinformation is getting harder to spot. Canadians are falling for it, expert warns

University of Alberta misinformation expert Timothy Caulfield on how tough it is to tell AI-generated doctors and health influencers apart from real ones.

Nearly nine in 10 Canadians turn to the internet for health information, according to a recent survey from the Canadian Medical Association (CMA), but many of them are encountering false or misleading claims along the way.

As AI-generated images and influencers become more sophisticated, experts warn the problem is poised to grow.

“It’s really remarkable. They look incredibly real,” said Timothy Caulfield, a professor of law and public health at the University of Alberta. “And keep in mind, it’s only going to get better. These images are only going to get more realistic.”

Caulfield told CTV News Channel Tuesday that research consistently shows people struggle to distinguish authentic content from AI-generated fakes — even when they believe they can.

“We think we can tell the difference. We can’t,” he said. “And very recent studies tell us that when AI images are included in content, we’re more likely to fall for misinformation.”

The issue is especially concerning in health care, where misleading claims can influence real-world decisions. Caulfield notes that AI-generated “influencers” are increasingly designed to appear relatable and trustworthy — sometimes styled as wellness coaches, spiritual figures or medical professionals — while promoting questionable supplements or treatments.

“There’s a real irony,” he said. “It’s a robot trying to look folksy and authentic in order to sell you stuff.”

In one example, Caulfield said even his wife, a physician, was briefly fooled by an AI-generated figure offering dubious thyroid advice.

“She was enraged,” he said. “What’s this monk doing talking about thyroid treatments?”

He warns that many of these accounts are not simply sharing opinions, but are monetizing attention.

“Be careful when they’re commodifying these messages,” he said. “A lot of these websites are selling something.”

The realism of AI imagery can make misinformation even more persuasive, particularly when viewed quickly on a phone screen.

“Imagine you’re in a hurry and you’re perhaps on the bus or on a train, and you look on your phone. It’s even more impressive,” he explained.

Still, he cautions against becoming overly cynical. The CMA survey also found that widespread misinformation is contributing to growing skepticism about legitimate medical advice.

According to the survey, the rise in false health information online has made 69 per cent of Canadians skeptical of any health information online, even from sources they think they should trust.

“There is good information online,” Caulfield said. “Be very, very careful when they’re actually giving health advice, especially if it’s something that is serious and it’s relevant to you.”

“Go see a licensed health professional,” he advised.

As AI tools continue to evolve, Caulfield says Canadians will need to adopt a mindset of digital humility.