Why do AI-savvy consumers harbor skepticism toward medical AI?
Patients who frequently use AI tend to readily trust AI-assisted diagnoses made by their physicians. Counterintuitively, however, those who would rank themselves among the very best-informed about AI tend to mistrust such diagnoses.
The paradoxical findings are from a survey-based study led by Catherine Chen, PhD, of Louisiana State University and Zhihan Cui, PhD, of Peking University and UCLA. The researchers surveyed 1,762 representative participants of varying demographics from around the U.S.
They offer three potential explanations for the surprising result described above:
1. The higher the self-reported AI knowledge, the more aware respondents may be of generative AI’s limitations, risks and ethical concerns.
“As general AI is not specifically designed for diagnostics, those with greater awareness may distrust its use in high-stakes contexts,” Chen and Cui surmise. More:
‘This heightened awareness can amplify concerns about reliability, accuracy and maturity in medical settings.’
2. The awareness of AI’s limitations may stem from perceived rather than actual risks.
People confident in their belief that health care AI is inadequate may misunderstand its true capabilities, the authors speculate. “Common concerns, such as AI’s rigidity and inability to personalize care, often drive aversion, although AI can sometimes outperform humans in these areas.”
‘Those who strongly believe in these misconceptions may report high AI knowledge while simultaneously exhibiting AI aversion.’
3. Overconfidence may skew individuals’ perceived levels of AI knowledge.
Self-reported AI knowledge may not reflect true AI literacy but rather an overestimation of one’s understanding coupled with an underestimation of AI’s capabilities, Chen and Cui remark. “Such overconfidence can lead to an exaggerated perception of AI’s weaknesses, contributing to the observed trust gap,” they write. “Unlike self-reported knowledge, experience reflects real interactions (e.g., using AI at work or in daily life).”
‘This practical experience likely offers a more accurate view of AI’s strengths and weaknesses, which may explain why frequent users showed less aversion.’
The Journal of Medical Internet Research published the study June 18. Read the whole thing.
—————————————
- In other research news:
- Regulatory:
- Funding:
Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.