Pew: Americans largely unsure about AI in healthcare

A Pew Research Center survey finds most Americans wary of generative AI in healthcare even as nearly a third have used AI chatbots for medical advice, with most of those users calling it helpful. More than 70% say it's very important that doctors disclose AI use, yet only 22% feel they actually understand how AI is used in their own care, and nearly half don't know if it's been used at all.

Researchers quoted in the piece note that open-source chatbots lack access to patient history and clinical context, making them prone to confidently wrong answers — false reassurance or needless panic, take your pick. Experts also flagged that chats with these free tools can be harvested to train future models, meaning the symptoms you typed in at 2 a.m. may live on as a data point.

Young adults, the group most distrustful of AI in general and least likely to have insurance, are also the most likely to turn to chatbots for medical advice. The piece notes that dedicated in-house medical AI tools, unlike the open-source variety, would actually be bound by HIPAA — which is either reassuring or a tacit admission of what's not currently covered.

The full dispatch is available from the source below.