Waltham, Mass. — U.S. patients are increasingly using artificial intelligence to research medications, diagnoses and other health information, but concerns over privacy, accuracy and human oversight remain major barriers to wider adoption, according to a new Wolters Kluwer Health survey.
The findings are part of the company’s 2026 Future Ready Healthcare survey and show sharp differences in confidence across age, gender and geography.
According to the survey, 74% of patients said they worry about the privacy of their personal health information when using AI tools, while 71% are concerned about their information being sold or leaked. Privacy concerns were higher among women, at 78%, compared with 65% of men.
Rural patients also reported greater concern about health data privacy, with 83% expressing worries compared with 72% of urban patients and 66% of suburban patients.
Despite those concerns, patients are using AI primarily for general medical research. More than half, or 54%, said they use AI to research medication side effects or drug interactions, while 52% use it to learn more about a diagnosis or condition. Another 43% use AI to simplify medical language or help explain test results.
The survey also found that 26% of patients sought medical care sooner after using AI to research a symptom or condition.
Concerns about inaccurate information remain widespread. Sixty-nine percent of respondents said they are worried about AI-generated answers being incorrect, with the figure approaching 80% among patients ages 25 to 29.
Human oversight remains important to most respondents. Eighty-nine percent said it is very or somewhat important for AI-generated health information to be validated by a human expert. In addition, 81% said AI tools used by doctors for clinical decisions should be formally approved by their health system, while 78% expect physicians to verify AI-generated information using other sources.
Liability was another major concern, with 75% of patients saying they are worried about who would be responsible if an AI-generated recommendation is wrong or causes harm.
Confidence in medical AI also varied significantly by gender and geography. Among men, 81% said they were extremely or somewhat confident in the reliability of AI-generated health answers, compared with 66% of women.
Confidence was highest among urban patients at 82%, followed by suburban patients at 72% and rural patients at 65%.
Patients were more comfortable with AI handling administrative functions than clinical decisions. Eighty-four percent of men and 74% of women said they were comfortable with AI autonomously scheduling appointments.
However, comfort declined when AI was used for diagnosis and treatment recommendations, with 61% of men and 36% of women saying they were comfortable with that level of autonomy.
Interest was also limited for several other uses. Only 30% said they would consider using AI to find a healthcare provider, 25% would use an AI chatbot for emotional support and 19% said they were comfortable using AI to navigate insurance coverage.
The 2026 Future Ready Healthcare survey was conducted by Ipsos from March 11 to March 14, 2026. It included 355 healthcare professionals, including 203 doctors and 152 nurses, as well as 254 U.S. patients.
