Dive Brief:
- Patients are less willing to use artificial intelligence for healthcare if it requires access to personal data or protected health information, according to a new survey.
- Healthcare consumers are primarily concerned about data privacy and the accuracy of AI-generated information, with around 89% saying AI responses in healthcare need human oversight, the survey by Wolters Kluwer shows.
- This trust gap means that less than a third of U.S. adults are using AI to find a healthcare provider, navigate their health insurance or create a diet plan, even though these applications are frequently marketed as key uses for AI in healthcare.
Dive Insight:
AI has quickly become integrated into Americans‘ lives, with 40% of survey respondents saying they use AI at least once daily in their personal lives, according to the survey, which was conducted by market research firm Ipsos on Wolters Kluwer’s behalf. The survey polled 254 U.S. patients in March.
AI remains a popular pathway for patients to better understand their health and treatment plans. About 1 in 4 survey respondents said they pursued healthcare sooner after using AI to research symptoms or conditions. Patients ages 25 to 34 and those living in rural areas also said the main benefit of AI was getting answers to health questions faster than going through the traditional route of waiting for answers from clinicians.
However, a fraction of those surveyed, about 30%, said they would use AI to find a healthcare provider, while only 28% would use AI to create a personal diet plan, workout routine or sleep schedule and 19% to navigate insurance coverage. This runs counter to prior research showing that patients are using AI to search for providers online and ask questions about health insurance.
Trust issues are the reason behind the reluctance to use AI for tasks involving personal data, the Wolters Kluwer survey states. Over 70% of patients worry about the privacy of their protected health information. These concerns are more pronounced among women and rural patients compared with their counterparts.
Not only that, but patients are also concerned about AI’s propensity for bias (72%) and hallucinations (69%). Younger patients aged 25 to 29 worry more about AI risks than those in other age groups.
Given these concerns, patients want human oversight of AI. A vast majority of respondents said it’s important that a human expert validate AI-generated health responses. Most also expect clinical AI tools to be formally approved by their health system. Still, it appears that AI tools are at least occasionally implemented without IT approval, which could worsen patient trust.
Communicating AI governance policies and creating standards for its use can help healthcare organizations build that trust. Patients highlighted concerns about the lack of clarity regarding who is responsible if AI recommendations are wrong or harmful. Nearly half of respondents said that AI tools used for health advice should undergo federal testing and approval, similar to scrutiny new medications face.
