Clinical experience and perception of risk affect the acceptance and trust of using AI in medicine.

Abstract

Background & aims: As Artificial Intelligence (AI) is progressively making inroads into clinical practice, questions have arisen as to whether acceptance of AI is skewed towards certain medical practitioner segments, even within particular specializations. This study aimed to examine distinct AI attitudes (including trust and acceptance) and intended behaviors among clinicians from contrasting backgrounds and levels of seniority/experience when interacting with AI.

Methods: Based on the results we divided participants into four groups, those who have (i) low experience and low risk perception, (ii) low experience and high risk perception, (iii) high experience and low risk perception, and (iv) high experience and perceived risk of AI use to be high. An ANCOVA model was constructed to test whether the four groups differ regarding their overall acceptance of AI.

Results: Data from 319 gastroenterologists show the presence of four distinct clusters of clinicians based upon experience levels and perceived risk typologies. Analysis of cluster-based responses further revealed that acceptance of AI was not uniform. Our findings showed that clinician experience and risk perspective have an interactive role in influencing AI acceptance. Senior clinicians with low-risk perception were highly accepting of AI, but those with high-risk perception of AI were substantially less accepting. In contrast, junior clinicians were more inclined to embrace AI when they perceived high risk, yet they hesitated to adopt AI when the perceived risk was minimal.

Conclusions: More experienced clinicians were more likely to embrace AI compared to their junior counterparts, particularly when they perceived the risk as low.