A new study reveals that American patients’ trust in health artificial intelligence is not keeping pace with its rapid spread in everyday life. The more a tool requires access to personal data or protected health information, the less willing patients are to use it, despite recognizing a clear benefit in gaining rapid access to information.
Ipsos conducted the survey for Wolters Kluwer in March, involving 254 patients in the United States. Forty percent of participants said they use artificial intelligence at least once a day in their personal lives, but its use for practical health tasks remained limited.
A Gap Between Benefit and Trust
About one-quarter of participants said that using AI to search for symptoms or health conditions prompted them to seek care earlier. Participants aged 25 to 34, as well as rural residents, also viewed speeding up access to health answers as the main benefit compared with waiting for a response from a physician.
However, fewer than one-third of participants expressed willingness to use AI in services based on personal data. About 30% said they might use it to find a healthcare provider, and 28% to create a personalized plan for diet, exercise, or sleep, while the figure was only 19% for understanding health insurance coverage.
Privacy Is Not the Only Concern
More than 70% of patients are concerned about the privacy of their protected health information. These concerns were more pronounced among women and rural residents. Likewise, 72% said they fear bias in AI systems, while 69% expressed concern that AI could produce fabricated or inaccurate answers, a problem known as “hallucination.” Participants between the ages of 25 and 29 were more concerned about AI risks than other age groups.
The findings indicate that patient acceptance is not linked merely to the existence of the tool, but to how it is used and who bears responsibility for its outcomes. About 89% believed that health answers generated by AI require human review, and most participants also expected clinical AI tools to receive formal approval from the healthcare system before being used.
What Does This Mean for Healthcare Organizations?
The study shows that healthcare institutions need publicly stated policies for AI governance and clear standards for its use, particularly when it handles protected health information. The use of tools without approval from the information technology department could also widen the trust gap rather than close it.
The most prominent open question is determining responsibility when an AI recommendation is wrong or harmful. Nearly half of participants said that AI tools used to provide health advice should be subject to testing and federal approval, similar to the scrutiny imposed on new drugs. The survey results alone do not establish that this regulatory model will be adopted, but they show that privacy, oversight, and accountability have become practical conditions for expanding the use of health artificial intelligence.