JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) has the potential to transform clinical practice and medical diagnostics. In times of workforce shortages, AI-based applications assist in clinical decision-making, patient monitoring, or administrative tasks. However, despite growing enthusiasm, integration into clinical practice remains limited due to concerns about usability, ethical implications, and physician acceptance. Addressing these challenges requires a deeper understanding of physicians’ attitudes toward AI. Engaging them in research and development can foster acceptance and improve adoption.

Objective:

To comprehensively assess physicians’ attitudes and expectations toward AI adoption in medicine.

Methods:

We conducted a mixed-methods study combining a web-based survey and qualitative interviews. The survey explored physicians’ perspectives on AI’s advantages and disadvantages, its role in medical decision-making, and its impact on patient-physician communication. Attitudes were measured using a 5-point Likert scale, covering both affective and cognitive dimensions. Exploratory factor analysis (EFA) identified underlying attitudinal factors, while Mann-Whitney U and Kruskal-Wallis tests examined differences in attitudes based on physicians’ age, medical specialty, AI familiarity, and more. Open-ended survey responses informed the development of the interview guide. 13 physicians, independent of the survey sample, participated in semi-structured interviews, which were analyzed using inductive coding and thematic analysis.

Results:

The survey yielded 498 valid responses. EFA revealed two factors: 1) “AI enthusiasm and acceptance” (Cronbach α=0.83) and 2) “AI skepticism and apprehension” (α=0.77). Physicians generally viewed AI positively (Median 4.00) and expressed lower skepticism (Median 3.62, reverse-coded, with higher scores indicating reduced concerns). Greater AI familiarity, research involvement, and AI use (both in daily life and professionally) were strongly associated with a more optimistic outlook and reduced skepticism. Physicians engaged in AI-related research reported significantly higher AI optimism (mean ranks: no AI research=54.32, AI research=111.52; P<.001) and lower skepticism (mean ranks: no AI research=70.45, AI research=108.27; P=.014). Those using AI professionally or willing to do so expressed similarly positive attitudes (mean ranks: no use=196.17, professional use=253.88; P=.001) and less skepticism (mean ranks: no use=218.86, plan to use=275.93; P=.001). Greater familiarity with AI tools was strongly associated with higher optimism (mean ranks: not familiar=169.86, very familiar=323.55; P<.001) and lower skepticism (mean ranks: not familiar=186.23, very familiar=296.90; P=.008). Chief physicians (mean rank: 277.32) were significantly less skeptical than assistant doctors (mean rank: 210.60; P=.014), while age and medical discipline did not influence attitudes. Qualitative analysis revealed six themes shaping AI adoption attitudes: 1) status quo, 2) AI dependency and negligence, 3) role changes and needs, 4) AI transparency and decision-making, 5) physician-patient relationship, and 6) framework for responsible AI integration. These findings informed a set of key propositions considered critical for successful AI adoption.

Conclusions:

AI in medicine is viewed positively overall, but attitudes are strongly shaped by direct experience and engagement rather than demographics. While concerns persist, they diminish with increased familiarity and professional use of AI. These findings highlight the need for targeted education, hands-on AI training, and efficient, standardized implementation strategies to enhance AI engagement and facilitate adoption.