Public still leery of AI chatbots in healthcare, misinformation

More than half of the American public don't trust AI chatbots in healthcare, saying in recent KFF polling that they don't think the tools can or will provide accurate medical information.

This comes as the healthcare industry balances the potential for AI to ease clinician workflows by supplementing patient-provider communication and its perennial patient trust issues.

"While most of the attention around AI in health is focused on how it can transform medical practice and create new business opportunities, consumers are also using it, and the jury is still out on whether it will empower or confuse them," KFF president and CEO Drew Altman said in a statement. "At KFF, our focus will be on how AI and other information technologies affect people."

Most AI chatbot users are unsure whether they themselves can parse fact from fiction when using the technologies to find health information. In total, only 9% of AI chatbot users think they can tell whether the information from a tool is true or false. A third said they are somewhat confident, while 35% said they are not too confident and 22% said they aren't at all confident.

Unsurprisingly, confidence in using AI chatbots is higher among younger users. While there aren't many people of any age who are totally confident they can pinpoint factual information in chatbots, a sizeable 43% of users ages 18 to 29 said they are somewhat confident. Among older users over age 65, that proportion was 23%.

Nevertheless, the public is using AI chatbots more and more. Around a third of respondents of any age said they use the tools at least once a week, and the KFF researchers predicted that share will grow as AI is integrated more seamlessly into search engines and social media platforms.

What's more, users are accessing AI chatbots to find healthcare information. Right now, about 17% of adults say they use chatbots to find information or advice about healthcare. That shakes out to a quarter of 18- to 29-year-olds and a fifth of those ages 30 to 49.

Again, despite growing utilization, users are apprehensive about the quality of information chatbots provide.

According to the poll, only 5% of adults are very confident that the information provided by AI chatbots for healthcare is accurate. While 31% said they are somewhat confident in the accuracy of information, a whopping 56% of users said they're not too confident or not confident at all that AI chatbots aren't peddling medical misinformation.

Notably, confidence in the accuracy of AI-generated health information is more common among regular chatbot users. While 38% of those who have used or interacted with an AI chatbot said they are at least somewhat confident in the accuracy of AI content, only 18% of nonusers said the same.

The jury's still out as to whether the public thinks AI chatbots will be a force of good or bad in the healthcare information space. Around a fifth of respondents said they think AI chatbots in healthcare will help people looking for health information, and 23% said they think it'll hurt. More than half (55%) said they weren't sure.

Again, patient trust is lacking among those who haven't used AI, older adults and Black individuals, KFF reported.

These findings echo previous surveys about consumer use of AI in healthcare. In May 2023, surveying from The University of Arizona Health Sciences showed that around half of patients don't fully trust AI-powered medical advice. Patients became more likely to trust the technology after AI got their clinician's seal of approval.

But there's also a lot of potential for AI in healthcare, especially in terms of helping patients get health information faster. AI deployed in EHRs and patient portals -- which should always get clinician review before being sent to patients -- can help patients get answers to their medical questions quickly and efficiently, some studies have shown.

Sara Heath has covered news related to patient engagement and health equity since 2015.