A Wolters Kluwer survey conducted by Ipsos, based on 254 U.S. patients in March, found that although 40% of respondents use AI at least once daily in their personal lives, many hesitate to use it for healthcare tasks that involve personal data. Fewer than a third said they would use AI to find a provider, navigate insurance, or create a diet plan — a finding that runs counter to earlier research suggesting patients already use AI for those purposes.
Privacy is the main sticking point. More than 70% of patients worry about the security of their protected health information, and concerns about AI bias and hallucinations are also widespread. Women and rural patients reported stronger privacy worries, while respondents aged 25 to 29 were more concerned about AI risks than other age groups. Still, about one in four said they sought care sooner after using AI to research symptoms, and younger and rural patients valued getting answers faster than waiting for clinicians.
Patients also want safeguards. Nearly 90% said AI-generated health responses need human validation, and most expect clinical AI tools to be formally approved by their health system. Almost half said such tools should undergo federal testing and approval. The survey notes that AI is sometimes deployed without IT approval, and that clearer governance and standards could help build trust, especially given patient concerns about who is responsible when AI recommendations are wrong or harmful.