More Americans Turn to Social Media and AI for Health Answers, and Experts Say That's Risky

More people are turning to social media and AI chatbots when a health question comes up, and a new study says that habit is starting to shape real medical decisions.

The study, published in the Journal of the American Medical Association, found that nearly 88% of U.S. adults surveyed use social media, and about 85% of those users have shared general or personal health information online. What stood out most to researchers, though, was how many people act on what they see.

"I think the biggest surprise that we found is that one in five people are actually making health related decisions based on what they saw on social media," said Rohan Khera, an associate professor at Yale University and a senior author of the study.

That's despite the fact that most people don't fully trust the content. About 78% of social media users said they believed health information on the platform was false or misleading, but many still let it guide their choices.

Khera says part of the problem is that convincing content isn't the same as accurate content. "It is so much easier as generative AI, like chatbots, have become around. You can make graphics, you can make images about anything, and that might be completely fake information," he said. "And then I think the other problem is there are people online whose entire revenue stream is the number of clicks they have."

AI chatbots aren't all bad, Khera says. They can help people understand medical terms or prepare questions before a doctor's visit. However, he warns the answer someone gets can depend on how they phrase the question, and a chatbot may miss warning signs a real doctor would catch.

"It will not keep helping you navigate to the highest risk conditions first, because it's not a triage tool," Khera said. "It's a personal assistant meant to have a conversation with you."