Physicians’ Attitudes Toward Artificial Intelligence in Medicine: Mixed Methods Survey and Interview Study
Open accessPublished version of a preprint
The DOI matches its Crossref record; 59 of the 59 references it cites were found in a registry or an index.
Abstract
Background: Artificial intelligence (AI) has the potential to transform clinical practice and diagnostics. Amid workforce shortages, AI-based applications assist in decision-making, patient monitoring, and administrative tasks. However, despite enthusiasm, integration into clinical practice remains limited because of concerns about usability, ethical implications, and physicians’ acceptance. Understanding physicians’ attitudes and engaging them in AI development may foster acceptance and adoption. Objective: This study aimed to comprehensively assess physicians’ attitudes toward AI in medicine. Methods: We conducted a mixed methods study combining a web-based survey and qualitative interviews. The survey explored physicians’ perspectives on the advantages and disadvantages of AI, its role in decision-making, and impact on physician-patient communication. Attitudes were measured using a 5-point Likert scale, covering cognitive and affective dimensions. An exploratory factor analysis (EFA) identified underlying attitudinal factors, while the Mann-Whitney U and the Kruskal-Wallis tests examined differences in attitudes based on physicians’ age, discipline, AI familiarity, and other variables. Overall, 13 physicians, independent of the survey sample, participated in semistructured interviews, which were analyzed using inductive coding and thematic analysis. Results: The survey yielded 498 responses. EFA revealed two factors: (1) AI enthusiasm and acceptance (Cronbach α=0.83) and (2) AI skepticism and apprehension (Cronbach α=0.77). Physicians reported high AI enthusiasm (median 4, IQR 3.57-4.29) and lower skepticism (median 3.62, IQR 3.20-4.20; reverse coded, with higher scores indicating reduced skepticism). Greater AI familiarity, use in daily life or professionally, and involvement in research were strongly associated with greater enthusiasm and lower skepticism. Physicians involved in AI-related research reported significantly higher enthusiasm (mean rank: AI research=111.52; no AI research=54.32; P<.001) and lower skepticism (mean rank: AI research=108.27; no AI research=70.45; P=.01). Those using AI professionally or intending to do so similarly expressed high enthusiasm (mean rank: professional use=253.88; no use=196.17; P=.001) and lower skepticism (mean rank: plan to use=275.93; no use=218.86; P=.001). Greater familiarity with AI tools was also strongly associated with higher enthusiasm (mean rank: very familiar=323.55; not familiar=169.86; P<.001) and lower skepticism (mean rank: very familiar=296.90; not familiar=186.23; P=.008). Chief physicians (mean rank 277.32) were significantly less skeptical than residents (mean rank 210.60; P=.01); however, age and discipline did not influence attitudes. The qualitative analysis identified six themes shaping physicians’ attitudes: (1) status quo, (2) AI dependency and negligence, (3) role changes and needs, (4) AI transparency and decision-making, (5) the physician-patient relationship, and (6) a framework for responsible AI integration. These findings led to several key propositions considered critical for AI adoption. Conclusions: AI in medicine is viewed positively, with attitudes shaped more by experience and engagement than by demographic factors. While concerns persist, they diminish with increased familiarity and professional use. These findings highlight the need for targeted education, hands-on training, and standardized implementation strategies to enhance AI engagement and facilitate adoption.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Similarly, physicians report both perceived benefits and concerns regarding AI in medical decision-making and patient-physician communication, underscoring that AI may be supportive but may also strain clinical relationships when trust, accountability, and AI literacy are insufficient [23].”
From Role-Differentiated AI Competencies and Curriculum Implications for Health Professions Education: Qualitative Study (2026) · cites this paper as [23]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Using thematic analysis in psychology
Braun, V., & Clarke, V. (2006) · Qualitative Research in Psychology · doi:10.1191/1478088706qp063oa · Cited by 194,414
Qualitative research & evaluation methods
Patton, M. Q. (2002) · Sage eBooks · Cited by 29,629
Standards for Reporting Qualitative Research: A Synthesis of Recommendations
O’Brien, B. C., Harris, I. B., Beckman, T. J., Reed, D. A., & Cook, D. A. (2014) · Academic Medicine · doi:10.1097/acm.0000000000000388 · Cited by 12,211
Comparison of convenience sampling and purposive sampling
Etikan, I. (2016) · American Journal of Theoretical and Applied Statistics · doi:10.11648/j.ajtas.20160501.11 · Cited by 11,412 · cited 2×
“Developed and administered with SoSci Survey (SoSci Survey GmbH), the survey was distributed using convenience sampling [26].”
Eysenbach, G. (2004) · Journal of Medical Internet Research · doi:10.2196/jmir.6.3.e34 · Cited by 6,480 · cited 1×
“Therefore, we developed a survey based on the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) guidelines [27] (supplementary materials 1 and 2 in Multimedia Appendix 1).”
Predicting the future - big data, machine learning, and clinical medicine
Obermeyer, Z., & Emanuel, E. J. (2016) · New England Journal of Medicine · doi:10.1056/nejmp1606181 · Cited by 3,652 · cited 1×
“Interestingly, contrary to common narratives [58], the physicians did not express fear of job replacement by AI.”
What is AI Literacy? Competencies and Design Considerations
Long, D., & Magerko, B. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376727 · Cited by 3,241 · cited 1×
“However, they expressed concern that increased patient access to generative AI tools (eg, ChatGPT and chatbots) might introduce misinformation into consultations [66].”
Rajpurkar, P., Chen, E., Banerjee, O., & Topol, E. J. (2022) · Nature Medicine · doi:10.1038/s41591-021-01614-0 · Cited by 3,212 · cited 2×
“In addition, racial bias has been reported in AI-based diagnostics [57].”
Using Mixed-Methods Sequential Explanatory Design: From Theory to Practice
Ivankova, N. V., Creswell, J. W., & Stick, S. L. (2006) · Field Methods · doi:10.1177/1525822x05282260 · Cited by 3,167 · cited 3×
“This study used an explanatory sequential mixed methods approach [24], consisting of (1) standardized surveys; (2) an open-ended survey question, “Is there anything else you would like to add to the topic of this survey?””
InfluentialPetty, R. E., Wegener, D. T., & Fabrigar, L. R. (1997) · Annual Review of Psychology · doi:10.1146/annurev.psych.48.1.609 · Cited by 2,290
Research works printed but not linked: 4
- Zapata, T, 2023, Eurohealth, 29, 6
- Rogers, EM, 2014, Diffusion of Innovations: An Integrated Approach to Communication Theory and Research
- Taherdoost, H, 2019, Int J Acad Res Manag, 8, 1
- Dresing, T, 2018, Praxisbuch Interview, Transkription & Analyse: Anleitungen und Regelsysteme für Qualitativ Forschende
Cited by
Park, Y., Han, Y., Kim, H., Ahn, S., Kron, F. W., & Lee, J. (2026) · JMIR Medical Education · doi:10.2196/97608
Metadata
- Authors
- Heinrichs, H., Kies, A., Nagel, S. K., & Kiessling, F.
- Affiliations
- RWTH Aachen University (DE)
- Fraunhofer Institute for Digital Medicine (DE)
- Corresponding author
- Kiessling, Fabian
- Venue
- Journal of Medical Internet Research
- Publisher
- JMIR Publications Inc.
- Published
- 2025-08-26
- Volume
- 27
- Pages
- e74187
- ISSN
- 1438-8871 (ISSN-L)
- PMC
- PMC12421205
- OpenAlex
- W4413682129
- Cited by
- 66
- References
- 59
- Citations by year
- 2026: 63 · 2025: 3
- FWCI
- 15.54
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Locations
- Journal of Medical Internet Research (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Fraunhofer-Publica (Fraunhofer-Gesellschaft) (submittedVersion)
- RWTH Publications (RWTH Aachen) (submittedVersion, other-oa)
- RWTH Publications (RWTH Aachen)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- artificial intelligence in healthcare, physician attitudes, AI acceptance, AI skepticism, AI familiarity, clinical decision-making, physician-patient communication, patient monitoring, clinical diagnostics, attitudinal factors
- MeSH terms
- Adult, Artificial Intelligence, Attitude of Health Personnel, Female, Humans, Interviews as Topic, Male, Middle Aged, Physicians, Surveys and Questionnaires
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.