CoChat
Sign Up
Cited by 46Artificial Intelligence in Healthcare and EducationInnovations in Medical EducationArtificial Intelligence in Education

Medical Students’ Attitudes Toward AI in Medicine and their Expectations for Medical Education

Open access
Kimmerle, Joachim, Timm, Jasmin, Festl-Wietek, Teresa, Cress, Ulrike, Herrmann-Werner, Anne·2023·Journal of Medical Education and Curricular Development
DOI 10.1177/23821205231219346PMC PMC10704950

Published version of a preprint

Open this paper in CoChat — free account

Abstract

Objectives: Artificial intelligence (AI) is used in a variety of contexts in medicine. This involves the use of algorithms and software that analyze digital information to make diagnoses and suggest adapted therapies. It is unclear, however, what medical students know about AI in medicine, how they evaluate its application, and what they expect from their medical training accordingly. In the study presented here, we aimed at providing answers to these questions. Methods: In this survey study, we asked medical students about their assessment of AI in medicine and recorded their ideas and suggestions for considering this topic in medical education. Fifty-eight medical students completed the survey. Results: Almost all participants were aware of the use of AI in medicine and had an adequate understanding of it. They perceived AI in medicine to be reliable, trustworthy, and technically competent, but did not have much faith in it. They considered AI in medicine to be rather intelligent but not anthropomorphic. Participants were interested in the opportunities of AI in the medical context and wanted to learn more about it. They indicated that basic AI knowledge should be taught in medical studies, in particular, knowledge about modes of operation, ethics, areas of application, reliability, and possible risks. Conclusions: We discuss the implications of these findings for the curricular development in medical education. Medical students need to be equipped with the knowledge and skills to use AI effectively and ethically in their future practice. This includes understanding the limitations and potential biases of AI algorithms by teaching the sensible use of human oversight and continuous monitoring to catch errors in AI algorithms and ensure that final decisions are made by human clinicians.

AI snapshot

Written from the paper’s full text.

TL;DR

In a survey of 58 medical students at one German university, participants were largely aware of AI in medicine and rated it as reliable, trustworthy, and technically competent but had low faith in it. They were highly interested in learning more and wanted basic AI knowledge—especially reliability, risks, ethics, modes of operation, and applications—integrated into medical education.

Key findings

  • Most participants (94.83%) were aware that AI is used in medicine; they associated AI mainly with algorithms (58.62%) and machine learning (48.28%), and cited diagnostics (86.21%) and surgeries (27.59%) as applications.
  • Participants rated AI in medicine as fairly reliable (M = 3.30, SD = 0.69), trustworthy (M = 3.58, SD = 0.71), and technically competent (M = 3.26, SD = 0.71), but had low faith in it (M = 2.34, SD = 0.71); they saw it as rather intelligent (M = 3.75, SD = 0.66) but not anthropomorphic (M = 1.99, SD = 0.64).
  • Interest in AI’s medical possibilities was high (M = 4.52, SD = 0.71), as was desire to learn more (M = 4.38, SD = 0.83) and to see AI addressed more in medical teaching (M = 4.17, SD = 0.92).
  • 86.21% agreed basic AI knowledge should be taught in medical studies; strongest support was for reliability (94.83%), possible risks (89.66%), modes of operation (77.59%), ethics (75.86%), and areas of application (75.86%). Reported problems included ethical concerns (53.45%), lack of control (43.10%), and potential lack of reliability (34.48%).

Methods

  • Online survey advertised in November 2022 via the medical student council e-mail distribution list at Eberhard Karls University, Tübingen, Germany; 58 medical or related-subject students completed the entire survey and consented to data use.
  • The questionnaire assessed awareness of AI in medicine, open-ended understanding and imagined applications, then provided a neutral definition of AI in medicine before measuring perceived reliability, technical competence, faith, trustworthiness, intelligence, and anthropomorphism using established scales on 5-point Likert or semantic differential formats.
  • It also asked about previous AI experience, whether basic AI knowledge should be taught, which specific aspects should be implemented in medical education, and open-ended perceived problems; results are reported descriptively as means, standard deviations, and percentages.

Limitations

  • The paper states that participants came only from one German university, reducing generalizability; future studies should include other medical schools.
  • The paper states that about one-fifth dropped out before finishing the survey, and compensation might reduce this in future studies.
  • The paper states that low faith in AI may reflect the specific item formulations, that no power calculation was made because no hypotheses were tested, and that some questions—previous AI experience and desired curricular aspects—had not been previously validated. Evident from the text: the sample is small and cross-sectional, and the analysis is descriptive without inferential testing.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1177/23821205231219346

Cites

10 of 10 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 5 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Participants referred primarily to diagnostics (86.21%) and surgeries (27.59%) as applications of AI in medicine.

    “As an application of AI in medicine, reference was made primarily to its use in diagnostics (86.21%) and surgeries (27.59%).”
    Quote in the paperNot a substantial claim
  • Findingown data

    Participants indicated only to a moderate extent that they experienced AI in a medical context (M = 2.69; SD = 1.43).

    “The participants indicated only to a moderate extent that they already had experience with AI (M = 2.85; SD = 1.41; on a 5-point scale), learned about AI in an educational context (M = 2.67; SD = 1.47), or experienced AI in a medical context (M = 2.69; SD = 1.43).”
    Quote in the paperNot a substantial claim
  • Findingown data

    There was less support for teaching technical basics (46.55%).

    “There was less support for teaching technical basics (46.55%), legal aspects (46.55%), future developments (46.55%), and current AI systems (43.10%).”
    Quote in the paperNot a substantial claim
  • Findingown data

    Participants indicated only to a moderate extent that they already had experience with AI (M = 2.85; SD = 1.41).

    “The participants indicated only to a moderate extent that they already had experience with AI (M = 2.85; SD = 1.41; on a 5-point scale), learned about AI in an educational context (M = 2.67; SD = 1.47), or experienced AI in a medical context (M = 2.69; SD = 1.43).”
    Quote in the paperNot a substantial claim
  • Findingown data

    There was less support for teaching legal aspects (46.55%).

    “There was less support for teaching technical basics (46.55%), legal aspects (46.55%), future developments (46.55%), and current AI systems (43.10%).”
    Quote in the paperNot a substantial claim
  • May combine two claims. 7
  • Classified as probably not a substantial claim. 7

  • May combine two claims.

    Findingown data

    Participants showed an adequate understanding of AI, referring to machine learning (48.28%), algorithms (58.62%), and neural networks (8.62%) as the most relevant aspects.

    “They showed an adequate understanding of AI, referring in the first open question to machine learning (48.28%), algorithms (58.62%), and neural networks (8.62%) as the most relevant aspects”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    The remaining 58 participants (35 females; mean age = 24.51 years, SD = 3.56 years) replied to all questions.

    “The remaining 58 participants (35 females; mean age = 24.51 years, SD = 3.56 years) replied to all questions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Seventeen of 84 participants dropped out before finishing the survey (dropout rate: 20.24%).

    “Eighty-four participants clicked on the link, but 17 dropped out before finishing the survey (dropout rate: 20.24%) and nine indicated not being medical students.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The selection of the specific scale format in each case followed the guidelines of the relevant research literature.

    “The selection of the specific scale format in each case followed the guidelines of the relevant research literature.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [9] Bartneck C Kulić D Croft E Zoghbi S. Measurement instruments for the anthropomorphism, animacy, likeability, perceived intelligence, and perceived safety of robots. Int J Soc Robot. 2009; 1: 71- 81. source

      Not checked in the free run.

    • cites [10] Lermann Henestrosa A Greving H Kimmerle J. Automated journalism: the effects of AI authorship and evaluative information on the perception of a science journalism article. Comp Hum Behav. 2023; 138: 107445.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Large parts of the survey used established questionnaires.

    “Finally, while large parts of the survey used established questionnaires, the questions about participants’ previous experience with AI and about the specific aspects that they would like to see implemented in medical education had not been previously validated.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Madsen and Gregor, 7

      Not checked in the free run.

    • cites Jian et al., 8

      Not checked in the free run.

    • cites Bartneck et al., 9

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Participants were particularly interested in ethical aspects of AI use and pointed out ethical concerns.

    “It is noteworthy that participants were particularly interested in ethical aspects of AI use and pointed out ethical concerns.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductreasoning

    No power calculation was made in advance because this was a survey study in which no hypotheses were tested.

    “Since this was a survey study in which no hypotheses were tested, no power calculation was made in advance.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    The questions about participants' previous experience with AI and about the specific aspects they would like to see implemented in medical education had not been previously validated.

    “Finally, while large parts of the survey used established questionnaires, the questions about participants’ previous experience with AI and about the specific aspects that they would like to see implemented in medical education had not been previously validated.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The low faith in AI finding could be due to the items of the faith scale being mainly aimed at situations in which the suitability of AI statements is unclear.

    “The items of this scale (see Appendix A) were mainly aimed at situations in which the suitability of the AI statements is unclear, so that these concrete item formulations could be responsible for this finding.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The study's participant pool consisted only of medical students from one German university, which may reduce the generalizability of the findings.

    “It is a limitation of this study that its participant pool consisted only of medical students from one German university, which may reduce the generalizability of our findings.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The findings indicate that medical students are very interested in AI in medicine and want to learn more about it in medical school.

    “The findings of this study indicate that medical students are very interested in AI in medicine and want to learn more about it in medical school.”
    Quote in the paperSubstantial claim
11 of 11

  • The paper relies on this without support.

    Not much is known about what prospective medical doctors know about AI and its application in medicine, how they assess this development, what they expect from their training, and what exactly they would like to see implemented in medical curricula.

    “At the same time, however, not much is known about what prospective medical doctors know about AI and its application in medicine, how they assess this development, what they expect from their training in this respect, and what exactly they would like to see implemented in medical curricula.”
    “It is unclear, however, what medical students know about AI in medicine, how they evaluate its application, and what they expect from their medical training accordingly.”
1 of 1

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?

  • Multiple articles have recommended various AI-related competencies or identified areas for learning about AI literacy in UGME.

    “Multiple articles have recommended various AI-related competencies or identified areas for learning about AI literacy in UGME [39, 74, 159, 160].”

    From Artificial intelligence in undergraduate medical education: an updated scoping review (2025) · cites this paper as [159]

  • Several ethical concerns regarding the use of AI in UGME are described, including algorithmic bias and safeguarding the privacy of students and patients.

    “Several ethical concerns regarding the use of AI in UGME are described [183]. There are concerns related to algorithmic bias [159] and safeguarding the privacy of students and patients [185, 186].”

    From Artificial intelligence in undergraduate medical education: an updated scoping review (2025) · cites this paper as [159]

  • 15 included studies were conducted in high-income countries and the remaining studies in low- and middle-income countries.

    “Specifically, 15 studies were conducted in HICs [26–40], while other studies were conducted in LMICs [23, 24, 41–49].”

    From Is it time to introduce an artificial intelligence curriculum in undergraduate medical education? medical students’ perspectives: a systematic review and meta-analysis (2026) · cites this paper as [28]

  • Literature searches located 7 survey research studies focusing on student or faculty AI perspectives.

    “Literature searches located 7 survey research studies focusing on student or faculty AI perspectives14–20; only three published AI training needs assessment studies were conducted in a U.S. medical school.15,19,21”

    From A Training Needs Analysis for AI and Generative AI in Medical Education: Perspectives of Faculty and Students (2025) · cites this paper as [20]

  • A lack of transparency and accountability in AI algorithms may hinder the understanding of their operational mechanisms.

    “A lack of transparency and accountability in AI algorithms may hinder the understanding of their operational mechanisms [159].”

    From Artificial intelligence in undergraduate medical education: an updated scoping review (2025) · cites this paper as [159]

    The Results (page unavailable) say students wanted to learn about AI’s modes of operation and identified lack of control as a concern. The Conclusions (page unavailable) call for understanding algorithmic limitations and human oversight. Neither passage, nor searches for transparency, accountability, and operational mechanisms, addresses whether a lack of transparency and accountability hinders understanding of how algorithms work. (LLM, conf=0.72)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. Madsen M, 2000, 11th Australasian Conference on Information Systems, 53, 6

Cited by

Sorted by citations

Authors
Kimmerle, J., Timm, J., Festl-Wietek, T., Cress, U., & Herrmann-Werner, A.
Affiliations
  • Leibniz-Institut für Wissensmedien (DE)
  • Tuebingen Institute for Medical Education, University of Tuebingen, Tuebingen, Germany
Corresponding author
Kimmerle, Joachim
ORCID
Venue
Journal of Medical Education and Curricular Development
Publisher
SAGE Publications
Published
2023-01
Volume
10
Pages
23821205231219346-23821205231219346
ISSN
2382-1205 (ISSN-L)
DOI
10.1177/23821205231219346
PMC
PMC10704950
OpenAlex
W4389444381
Cited by
46
References
10
Citations by year
2026: 9 · 2025: 23 · 2024: 13 · 2023: 1
FWCI
1.84
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Medical Education and Curricular Development (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
German Federal Ministry of Education and Research
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Innovations in Medical Education
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
Keywords
artificial intelligence in healthcare, medical students' attitudes, AI literacy, medical curriculum, AI ethics, trust in AI, AI reliability, algorithmic bias, human oversight of AI

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

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