The DOI matches its Crossref record; 75 of the 75 references it cites were found in a registry or an index; 119 of the 161 sources its claims cite were checked for support.
Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review
Open accessPublished version of a preprint
Abstract
Background: Generative artificial intelligence (AI) technologies are increasingly being utilized across various fields, with considerable interest and concern regarding their potential application in medical education. These technologies, such as Chat GPT and Bard, can generate new content and have a wide range of possible applications. Objective: This study aimed to synthesize the potential opportunities and limitations of generative AI in medical education. It sought to identify prevalent themes within recent literature regarding potential applications and challenges of generative AI in medical education and use these to guide future areas for exploration. Methods: We conducted a scoping review, following the framework by Arksey and O'Malley, of English language articles published from 2022 onward that discussed generative AI in the context of medical education. A literature search was performed using PubMed, Web of Science, and Google Scholar databases. We screened articles for inclusion, extracted data from relevant studies, and completed a quantitative and qualitative synthesis of the data. Results: Thematic analysis revealed diverse potential applications for generative AI in medical education, including self-directed learning, simulation scenarios, and writing assistance. However, the literature also highlighted significant challenges, such as issues with academic integrity, data accuracy, and potential detriments to learning. Based on these themes and the current state of the literature, we propose the following 3 key areas for investigation: developing learners’ skills to evaluate AI critically, rethinking assessment methodology, and studying human-AI interactions. Conclusions: The integration of generative AI in medical education presents exciting opportunities, alongside considerable challenges. There is a need to develop new skills and competencies related to AI as well as thoughtful, nuanced approaches to examine the growing use of generative AI in medical education.
Overview
CoChat Summary
TL;DR
This scoping review synthesizes 41 articles published in 2023 on generative AI in medical education, identifying opportunities such as self-directed learning, simulation, and writing assistance alongside challenges including academic integrity, accuracy/hallucination, and potential learning harms. It proposes three future investigation areas: learners’ critical AI evaluation skills, rethinking assessment, and studying human-AI interaction.
Key findings
- From 2761 unique titles, after removing 168 duplicates, 2593 studies were screened; 2425 were unrelated and 127 were excluded, yielding 41 included articles. Of these, 21/41 (51.2%) were opinion pieces and 20/41 (48.8%) were original research; 16 original research articles reported generative AI performance on standardized assessments. All included studies were published in 2023.
- Thematic analysis identified potential benefits in test performance/preparation, novel learning strategies (personalized learning plans/materials, feedback, communication skills training, clinical image generation, medical humanities exercises), and writing/research assistance (assisting non-native speakers, translations, literature review/summarization).
- Challenges included academic integrity concerns (cheating, technological plagiarism, need for policy and disclosure guidance), accuracy/dependability issues (reliance on training data, lack of up-to-date information, hallucination, unwarranted confidence, misinformation propagation, limited accuracy in specific areas), and potential detriments to learning (overdependence, assessment challenges, propagating inaccurate information, inequities in access).
- The authors propose three key areas for investigation: developing learners’ AI literacy/critical appraisal and navigating uncertainty, considering implications for assessment, and exploring human-AI interaction.
Methods
- Scoping review following the Arksey and O’Malley framework; searched PubMed, Web of Science, and Google Scholar for English-language articles published from January 1, 2022, to June 21, 2023, discussing generative AI in medical education.
- The first 100 articles were independently screened by both authors based on titles and abstracts, yielding substantial agreement (Cohen kappa=0.76); one author screened the remaining studies, and a random subset of full-text articles was independently reviewed by the second author. Conflicts were resolved through discussion and consensus.
- Data abstraction used a structured form; descriptive statistics summarized study demographics, and qualitative data underwent thematic analysis guided by Braun and Clarke, with themes and subthemes reported in Table 2.
Limitations
- Stated: the search was restricted to English-language articles and to publications from 2022 onward, excluding earlier related literature such as natural language processing in medical education; thematic analysis has inherent subjectivity; the literature is rapidly evolving; and the lack of empirical studies limits definitive conclusions, with most benefits and challenges remaining speculative.
- Evident: only 1 study examined learner attitudes toward generative AI in medical education, and no curricula specifically focusing on generative AI were found; most included articles were opinion pieces or speculative, limiting empirical grounding.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures


Verification
Claims the paper makes
33 Valid · 2 Not valid · 10 Not checked · 48 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“For instance, Webb [16] discussed the potential for generative AI to enhance communication skills for emergency medicine physicians, particularly for delivering difficult news. This was achieved by simulating patient reactions and dialogues during the disclosure of a new cancer diagnosis [16].”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Potential Benefits of Generative AI in Medical Education
What the sources say
- Supportedcites [16] source
Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Numerous studies underscored the potential of these AI models to adapt to individual learners' requirements, offering a customized learning experience [1,14,15,20,22,34].”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Potential Benefits of Generative AI in Medical Education
What the sources say
- Partly supportedcites [1] sourceExperimental — accuracy not yet measured
In The Interview, ChatGPT says language models can analyze individual students’ performance and generate materials addressing their weaknesses, supporting the claimed potential for customized learning. This paper presents that point as a possibility in a conversation, not as a finding established by numerous studies. (LLM, conf=0.86)
- Supportedcites [14] source
The Potential Benefits section says generative AI can analyze each student's strengths and weaknesses to create personalized assessments and help educators develop learning plans addressing individual needs. The Conclusions likewise describe GLMs generating individualized content for students. These passages support the attributed potential for customized learning. (LLM, conf=0.92)
- Supportedcites [15] source
In “Potential Applications and Benefits of ChatGPT in Medical Education” (page not provided), the paper explicitly describes ChatGPT’s potential for personalized education. It says ChatGPT can respond at an appropriate level and explain difficult concepts in ways that adapt to a struggling student’s level. The paper notes that tracking progress and adjusting teaching style are not yet developed; the citing sentence claims potential, not an established capability. (LLM, conf=0.91)
- Supportedcites [20] source
The ‘ChatGPT Role in Medical education’ section says ChatGPT can provide personalized study plans and learning materials suited to students’ different learning styles and abilities. This supports the cited claim about its potential to offer customized learning. (LLM, conf=0.93)
- Supportedcites [22] source
The Opportunities section, under “Personalized Study Plans and Learning Materials” (page not provided), says students can use LLMs to generate study plans tailored to their strengths, weaknesses, goals, and preferences. It also describes personalized materials targeting areas where a student needs improvement, supporting the attributed potential for customized learning. (LLM, conf=0.92)
- Supportedcites [34] source
The Discussion says ChatGPT’s major advantage for medical students is personalized learning suited to their individual needs. That supports this source’s contribution to the citing sentence’s claim about customized learning. (LLM, conf=0.90)
plainstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“As touched upon in the preceding paragraph, a significant worry cited by numerous authors is the potential threat to academic integrity and the possible misuse of this technology [12,14,15,19,20,22,23,25,40-43].”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Potential Limitations of Generative AI in Medical Education
What the sources say
- Supportedcites [12] source
The “ChatGPT in medical education” section warns that students could exploit ChatGPT to write entire assignments rather than use it to refine their work. This supports citing the paper as one of the authors raising concerns about misuse and academic integrity; the supplied copy gives no page number. (LLM, conf=0.91)
- Supportedcites [14] source
The Introduction identifies misuse of AI-generated content and academic integrity as concerns in medical education. The Challenges and Ethical Considerations section calls AI-enabled academic dishonesty a critical issue and gives examples of how generative AI tools could be misused. (LLM, conf=0.94)
- Supportedcites [15] source
In “ChatGPT’s responses,” Lee identifies student reliance on ChatGPT as a concern that may lead to academic dishonesty or cheating, including using it to complete assignments. The Figure 1 discussion also says that ChatGPT’s use for academic misconduct calls traditional assessments into question. (LLM, conf=0.92)
- Supportedcites [19] source
In “Ethics and Integrity in Academic Work,” the authors explicitly discuss the risk that AI technologies could be misused and ask how their use affects professional integrity and accountability. The abstract likewise identifies ethical concerns about LLMs in health professions education scholarship. This supports the concern attributed to this source, without requiring it alone to establish what numerous authors say. (LLM, conf=0.92)
- Supportedcites [20] source
The “ChatGPT Role in Medical education” section explicitly says ChatGPT has raised concerns about plagiarism and cheating. These are the academic-integrity risks the citing sentence describes; the claim that numerous authors raise the concern refers to its collection of citations, not to this source alone. (LLM, conf=0.91)
- Supportedcites [22] source
The Introduction identifies plagiarism among the serious challenges of adopting LLMs in medical education. The Academic Dishonesty section explains how LLMs could be misused for cheating and plagiarism, directly supporting the concern attributed to this source. (LLM, conf=0.94)
- Not checked — no full text of the sourcecites [23] source
- Not checked yetcites [25] doi:10.6145/jme.202303_27(1).0005
- Supportedcites [40] source
The paper’s main text warns that students may use ChatGPT to cheat and calls for surveillance to prevent its use in medical assignments. It also identifies ethical concerns and the potential for misuse, supporting the concern attributed to this source. (LLM, conf=0.92)
- Partly supportedcites [41] sourceExperimental — accuracy not yet measured
The Abstract warns that plagiarism remains a concern when ChatGPT is used for academic writing, and the Academic writing section reports plagiarism matches in a generated literature review. These passages support a concern about academic integrity, but the paper’s medical education section does not discuss misuse or academic-integrity threats in that setting. (LLM, conf=0.82)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Several studies focused on the role of generative AI models in tests of medical knowledge [8-11,13,26,27,31-39].”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Potential Benefits of Generative AI in Medical Education
What the sources say
- Supportedcites [8] source
The Introduction (page unavailable) says the study evaluates ChatGPT, a large language model, using questions from the USMLE. The Results report its accuracy on those exam questions, confirming that this paper is an example of a study testing a generative AI model on medical knowledge. (LLM, conf=0.94)
- Supportedcites [9] source
The section on ChatGPT’s performance on exams describes the European Exam in Core Cardiology as a knowledge-based assessment and states that the authors evaluated ChatGPT on it. This paper is therefore an example of a study testing a generative AI model on medical knowledge. (LLM, conf=0.94)
- Supportedcites [10] source
The Conclusions report that this study tested ChatGPT on higher-order medical biochemistry questions, including multiple-choice questions. It is therefore an example of a study examining a generative AI model on a test of medical knowledge. (LLM, conf=0.92)
- Supportedcites [11] source
This paper is one of the studies described in the citing sentence. Its Results section evaluates GPT-3.5 and GPT-4 on 280 Korean general surgery board-exam questions, a test of medical knowledge. (LLM, conf=0.94)
- Supportedcites [13] source
The Discussion states that this study tested ChatGPT on Taiwan’s Family Medicine Board Exam, making it an example of a study evaluating a generative AI model on medical knowledge. The Introduction also describes other studies testing ChatGPT on medical examinations. (LLM, conf=0.93)
- Supportedcites [26] source
The Abstract states that the study assessed ChatGPT’s accuracy on the Plastic Surgery Inservice Training Examination using questions from the 2022 exam. This paper therefore exemplifies a study of a generative AI model on a test of medical knowledge. (LLM, conf=0.94)
- Not checked — no full text of the sourcecites [27] source
- Supportedcites [31] source
The Abstract describes the authors testing ChatGPT on a clinical toxicology vignette and evaluating its answers to medical questions. That case-based assessment supports citing this paper as a study of a generative AI model's medical knowledge; the paper does not claim to have administered a standardized examination. (LLM, conf=0.86)
- Supportedcites [32] source
The Background identifies ChatGPT as an AI chatbot and says the study examines how accurately it can answer medical-examination questions. The Conclusion reports its performance on a parasitology examination, so this paper is an example of a study of generative AI in a medical-knowledge test. Page numbers are not provided in our copy. (LLM, conf=0.92)
- Supportedcites [33] source
The Introduction states that the study aimed to assess ChatGPT’s performance on examinations of medical knowledge, including USMLE Step 1 and Step 2. This paper is therefore an example of the studies described in the citing sentence. Page numbers were not supplied in our copy. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Several authors underscore that the knowledge base of these models is constrained by their training data, given that most models lack internet access to retrieve the most current information [10,22,34,37,44].”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Potential Limitations of Generative AI in Medical Education
What the sources say
- Partly supportedcites [10] sourceExperimental — accuracy not yet measured
The Discussion (page not available) says ChatGPT’s information dataset ended in 2021 and that its answers may omit recent advances; the Conclusions likewise link weaker performance on recent topics to training-database limitations. The paper does not say that most generative AI models lack internet access, so it does not support that explanation or its extension to most models. (LLM, conf=0.86)
- Partly supportedcites [22] sourceExperimental — accuracy not yet measured
The “Challenges > Limited Knowledge” section says LLMs depend on training data that may omit the latest medical knowledge and that most LLMs have a static knowledge base. It does not say that most models lack internet access or identify that as the reason they cannot retrieve current information. (LLM, conf=0.87)
- Partly supportedcites [34] sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) suggests that limitations in ChatGPT’s training may contribute to its uneven performance across microbiology topics. It does not say that most generative AI models lack internet access or that this prevents them from retrieving current information. (LLM, conf=0.81)
- Partly supportedcites [37] sourceExperimental — accuracy not yet measured
The Methods section ('ChatGPT Is an LLM'; page unavailable) describes ChatGPT as trained on internet text, and the Discussion links its performance to the availability and quality of training data. The paper does not say that most models lack internet access or cannot retrieve current information, so that explanation is unsupported by this source. (LLM, conf=0.85)
- Not checked — no full text of the sourcecites [44] source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
30 more claims · 9 with their cited sources checked
Create a free account to see all 93 claims and to verify the 11 references still open.
- Probably not a substantial claim. 39
- It may state two separate claims as one. 25
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?
The cited source doesn’t support the claim as written.
In the citing paper
“studies before 2023 were constrained by the technology’s maturity, missing the recent shift from theoretical exploration to operational implementation [1]”
From Applications, Challenges, and Prospects of Generative Artificial Intelligence Empowering Medical Education: Scoping Review (2025) · cites this paper as [1]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“GAI rapidly expands into diverse applications, enabling content generation across text, images, and audio through the analysis of extensive datasets [1].”
From Applications, Challenges, and Prospects of Generative Artificial Intelligence Empowering Medical Education: Scoping Review (2025) · cites this paper as [1]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Additional factors might include insufficient training data for the LLM [17,35,36], context misinterpretation, lack of external validation, and overreliance on pattern recognition, all of which could contribute to variable degrees of hallucination, depending on the specific LLM and prompt structure.”
From Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study (2024) · cites this paper as [36]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Several scoping reviews have examined AI within medical education23–28”
From What the AI era doctor should know: a scoping review of proposed artificial intelligence competencies for medical education (2026) · cites this paper as [28]
The cited source says what the claim attributes to it.
In the citing paper
“but these syntheses largely describe broad “AI literacy” themes or program-level content rather than systematically extracting and organizing competency- and learning-objective-level outcomes for undergraduate medical education (UME).”
From What the AI era doctor should know: a scoping review of proposed artificial intelligence competencies for medical education (2026) · cites this paper as [28]
2 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Study Characteristics
“Our initial search identified 2761 unique titles (Figure 1).”
§ Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review > Results > Study Characteristics
“After removing 168 duplicates, 2593 studies were available for screening.”
Citations
Cites
Using thematic analysis in psychology
Braun, V., & Clarke, V. (2006) · Qualitative Research in Psychology · doi:10.1191/1478088706qp063oa · Cited by 194,414
Scoping studies: towards a methodological framework
Arksey, H., & O'Malley, L. (2005) · International Journal of Social Research Methodology · doi:10.1080/1364557032000119616 · Cited by 36,742
Why Most Published Research Findings Are False
Ioannidis, J. P. A. (2005) · PLoS Medicine · doi:10.1371/journal.pmed.0020124 · Cited by 10,505
High-performance medicine: the convergence of human and artificial intelligence
Topol, E. J. (2019) · Nature Medicine · doi:10.1038/s41591-018-0300-7 · Cited by 10,450 · cited 1×
“The tensions of navigating information provided by technology and other sources are fertile ground for exploration and discussion among learners, particularly as AI begins to drive more clinical decisions [4].”
BackgroundWhy Most Published Research Findings Are False
Ioannidis, J. P. A. (2005) · CHANCE · doi:10.1080/09332480.2005.10722754 · Cited by 4,702 · cited 1×
“Considering the primary medical literature, most published research findings are suggested to be false [55].”
BackgroundKung, T. H., Cheatham, M., Medenilla, A., Sillos, C., De Leon, L., Elepaño, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., Maningo, J., et al. (2023) · PLOS Digital Health · doi:10.1371/journal.pdig.0000198 · Cited by 4,115 · cited 8×
“Only a handful of these studies delved into the potential implications of generative AI's performance on these tests [8,33].”
BackgroundInfluentialKung, T. H., Cheatham, M., ChatGPT, Medenilla, A., Sillos, C., De Leon, L., Elepaño, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., et al. (2022) · doi:10.1101/2022.12.19.22283643 · Cited by 4,115
The answer is 17 years, what is the question: understanding time lags in translational research
Morris, Z. S., Wooding, S., & Grant, J. (2011) · Journal of the Royal Society of Medicine · doi:10.1258/jrsm.2011.110180 · Cited by 3,340 · cited 1×
“Medical textbooks, often considered the gold standard in medical knowledge, are perpetually outdated, are limited in scope, and may contain inaccuracies [52-54].”
BackgroundMorgan, D. L. (2007) · Journal of Mixed Methods Research · doi:10.1177/2345678906292462 · Cited by 2,503 · cited 1×
“Data uncertainty can be viewed from a positivist perspective with error margins and reliability estimates or from a pragmatic perspective, which focuses on the data’s utility [50].”
BackgroundGebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé, H., & Crawford, K. (2018) · arXiv · doi:10.48550/arxiv.1803.09010 · Cited by 2,474
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., III, H. D., & Crawford, K. (2021) · Communications of the ACM · doi:10.1145/3458723 · Cited by 2,469
Gilson, A., Safranek, C. W., Huang, T., Socrates, V., Chi, L., Taylor, R. A., & Chartash, D. (2023) · JMIR Medical Education · doi:10.2196/45312 · Cited by 2,078 · cited 1×
“Only a handful of these studies delved into the potential implications of generative AI's performance on these tests [8,33].”
BackgroundArtificial hallucinations in ChatGPT: implications in scientific writing
Alkaissi, H., & McFarlane, S. I. (2023) · Cureus · doi:10.7759/cureus.35179 · Cited by 1,540 · cited 4×
“[28] Confidence expressed by models “Further, AI-generated content can potentially produce misinformation or biased information...” [14] Misinformation propagation “Consequently, the current level of accuracy is not yet sufficient for immediate clinical application in patient care.””
BackgroundInfluentialChatGPT is fun, but not an author
Thorp, H. H. (2023) · Science · doi:10.1126/science.adg7879 · Cited by 1,437
Gao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2022) · doi:10.1101/2022.12.23.521610 · Cited by 1,099 · cited 1×
“Instances of technological plagiarism already exist, in which AI has generated abstracts or entire scientific papers with minimal human involvement [66,67].”
BackgroundGao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2023) · npj Digital Medicine · doi:10.1038/s41746-023-00819-6 · Cited by 1,099
ChatGPT listed as author on research papers: many scientists disapprove
Stokel-Walker, C. (2023) · Nature · doi:10.1038/d41586-023-00107-z · Cited by 868 · cited 1×
“Instances of technological plagiarism already exist, in which AI has generated abstracts or entire scientific papers with minimal human involvement [66,67].”
BackgroundEysenbach, G. (2023) · JMIR Medical Education · doi:10.2196/46885 · Cited by 840 · cited 7×
“[1] Creation of learning materials “By serving as a virtual teaching assistant, ChatGPT could be leveraged to provide students with real-time and personalized feedback.””
BackgroundInfluentialLarge Language Models in Medical Education: Opportunities, Challenges, and Future Directions
Abd-alrazaq, A., AlSaad, R., Alhuwail, D., Ahmed, A., Healy, P. M., Latifi, S., Aziz, S., Damseh, R., Alabed Alrazak, S., & Sheikh, J. (2023) · JMIR Medical Education · doi:10.2196/48291 · Cited by 722 · cited 10×
“[22] Cheating on examinations “Student dependency on the language model may also propagate academic dishonesty or ‘cheating.”
BackgroundInfluentialAI-Mediated Communication: Definition, Research Agenda, and Ethical Considerations
Hancock, J. T., Naaman, M., & Levy, K. (2020) · Journal of Computer-Mediated Communication · doi:10.1093/jcmc/zmz022 · Cited by 658 · cited 1×
“An emerging field of AI-mediated communication focuses on AI’s influence on our interactions with others [68].”
BackgroundChatGPT - Reshaping medical education and clinical management
Khan, R. A., Jawaid, M., Khan, A. R., & Sajjad, M. (2023) · Pakistan Journal of Medical Sciences · doi:10.12669/pjms.39.2.7653 · Cited by 628 · cited 10×
“The precision and trustworthiness of generative AI are fundamental concerns thoroughly elaborated in many publications [8,11-15, 20,22, 24, 32, The propensity of these systems to generate and propagate misinformation is a notable risk.”
BackgroundInfluentialBoscardin, C. K., Gin, B., Golde, P. B., & Hauer, K. E. (2024) · Academic Medicine · doi:10.1097/acm.0000000000005439 · Cited by 620 · cited 1×
“Within health profession education, AI literacy encompasses understanding the capabilities of AI; integrating AI into practice; and ensuring inclusion, equity, and responsible use of AI [48].”
BackgroundThe rise of ChatGPT : Exploring its potential in medical education
Lee, H. (2024) · Anatomical Sciences Education · doi:10.1002/ase.2270 · Cited by 596 · cited 10×
“The precision and trustworthiness of generative AI are fundamental concerns thoroughly elaborated in many publications [8,11-15, 20,22, 24, 32, The propensity of these systems to generate and propagate misinformation is a notable risk.”
BackgroundInfluentialEvaluating the Performance of ChatGPT in Ophthalmology
Antaki, F., Touma, S., Milad, D., El-Khoury, J., & Duval, R. (2023) · Ophthalmology Science · doi:10.1016/j.xops.2023.100324 · Cited by 563 · cited 1×
“A significant issue lies in data accuracy, with many authors drawing attention to this technology’s propensity to “hallucinate,” or create false information, and its knowledge being confined to the training data set [1,10,21,22,25,27,28,34,37,44].”
BackgroundAccuracy of a Generative Artificial Intelligence Model in a Complex Diagnostic Challenge
Kanjee, Z., Crowe, B., & Rodman, A. (2023) · JAMA · doi:10.1001/jama.2023.8288 · Cited by 449
Curriculum Development for Medical Education
Thomas, P. (2016) · Johns Hopkins University Press eBooks · doi:10.1353/book.44600 · Cited by 415
Artificial intelligence in medical education: a cross-sectional needs assessment
Civaner, M. M., Uncu, Y., Bulut, F., Chalil, E. G., & Tatli, A. (2022) · BMC Medical Education · doi:10.1186/s12909-022-03852-3 · Cited by 363
The future of medical education and research: Is ChatGPT a blessing or blight in disguise?
Arif, T. B., Munaf, U., & Ul-Haque, I. (2023) · Medical Education Online · doi:10.1080/10872981.2023.2181052 · Cited by 317 · cited 7×
“Several articles voice concerns about potential academic dishonesty [12,14,15,19,20,22,23,25,40-43].”
BackgroundInfluentialEthical use of Artificial Intelligence in Health Professions Education: AMEE Guide No. 158
Masters, K. (2023) · Medical Teacher · doi:10.1080/0142159x.2023.2186203 · Cited by 302 · cited 6×
“Several articles voice concerns about potential academic dishonesty [12,14,15,19,20,22,23,25,40-43].”
BackgroundInfluentialHuh, S. (2023) · Journal of Educational Evaluation for Health Professions · doi:10.3352/jeehp.2023.20.1 · Cited by 278 · cited 2×
“Despite the remarkable performance of these models on standardized tests, they still commit significant errors, and their performance is often on par with that of novice learners [32,35,36].”
Background(2015) · Choice Reviews Online · doi:10.5860/choice.189890 · Cited by 264
The Advent of Generative Language Models in Medical Education
Karabacak, M., Ozkara, B. B., Margetis, K., Wintermark, M., & Bisdas, S. (2023) · JMIR Medical Education · doi:10.2196/48163 · Cited by 208 · cited 10×
“[28] Confidence expressed by models “Further, AI-generated content can potentially produce misinformation or biased information...” [14] Misinformation propagation “Consequently, the current level of accuracy is not yet sufficient for immediate clinical application in patient care.””
BackgroundInfluentialThe Butterfly Effect in artificial intelligence systems: Implications for AI bias and fairness
Ferrara, E. (2024) · Machine Learning with Applications · doi:10.1016/j.mlwa.2024.100525 · Cited by 208
Oh, N., Choi, G.-S., & Lee, W. Y. (2023) · Annals of Surgical Treatment and Research · doi:10.4174/astr.2023.104.5.269 · Cited by 201 · cited 8×
“The precision and trustworthiness of generative AI are fundamental concerns thoroughly elaborated in many publications [8,11-15, 20,22, 24, 32, The propensity of these systems to generate and propagate misinformation is a notable risk.”
BackgroundInfluentialMoldt, J.-A., Festl-Wietek, T., Madany Mamlouk, A., Nieselt, K., Fuhl, W., & Herrmann-Werner, A. (2023) · Medical Education Online · doi:10.1080/10872981.2023.2182659 · Cited by 186 · cited 3×
“[29] Need for further training in limitations Theme 6: Potential detriments to learning “Lastly, there is a need to delve deeper into the possible consequences of overdependence on LLMs in medical education.””
BackgroundInfluentialEvaluation in health professions education—Is measuring outcomes enough?
Allen, L., Hay, M., & Palermo, C. E. (2021) · Medical Education · doi:10.1111/medu.14654 · Cited by 166 · cited 1×
“Curriculum evaluation methodologies focusing on context, such as the Context, Input, Process, Product (CIPP) model, theory-driven evaluation, or realist evaluation, might be particularly adept at accounting for and examining human-AI interaction within an educational intervention [65].”
BackgroundEarly applications of ChatGPT in medical practice, education and research
Sedaghat, S. (2023) · Clinical Medicine · doi:10.7861/clinmed.2023-0078 · Cited by 163
Humar, P., Asaad, M., Bengur, F. B., & Nguyen, V. (2023) · Aesthetic Surgery Journal · doi:10.1093/asj/sjad130 · Cited by 155 · cited 2×
“A few studies outline potential scenarios in which educators or learners might interact with AI systems, such as in self-directed learning, simulation environments, and writing assistance [8,11,12,14,19,22,23,25,28,34,35,38,40,64].”
BackgroundGottlieb, E. R., Ziegler, J., Morley, K., Rush, B., & Celi, L. A. (2022) · JAMA Internal Medicine · doi:10.1001/jamainternmed.2022.2587 · Cited by 134 · cited 1×
“Lack of representation and bias in medical records data are major concerns, and we are only beginning to recognize biases in technology that has been in use in health care for years [57-59].”
BackgroundChatGPT takes on the European Exam in Core Cardiology: an artificial intelligence success story?
Skalidis, I., Cagnina, A., Luangphiphat, W., Mahendiran, T., Muller, O., Abbe, E., & Fournier, S. (2023) · European Heart Journal - Digital Health · doi:10.1093/ehjdh/ztad029 · Cited by 128 · cited 1×
“[9] Specialty exam performance “It can be concluded that ChatGPT helps in seeking answers for higher-order reasoning questions in medical bio-chemistry.””
BackgroundPushing Critical Thinking Skills With Multiple-Choice Questions: Does Bloom’s Taxonomy Work?
Zaidi, N. L. B., Grob, K. L., Monrad, S. U., Kurtz, J. B., Tai, A. W., Ahmed, A. Z., Gruppen, L. D., & Santen, S. A. (2017) · Academic Medicine · doi:10.1097/acm.0000000000002087 · Cited by 128 · cited 1×
“For curriculum evaluation–based assessments, educators often aim to test learners’ capabilities to comprehend and perform higher-order cognitive skills [60,61].”
BackgroundBreaking into the black box of artificial intelligence
Savage, N. (2022) · Nature · doi:10.1038/d41586-022-00858-1 · Cited by 124 · cited 1×
“Moreover, concerns have arisen that generative AI may produce biased content or lack representation of all populations [8,11-15, 19, 20, 22, 24, This inability to observe how a response is generated is often referred to as “the black box” problem [49].”
BackgroundChatGPT in medical school: how successful is AI in progress testing?
Friederichs, H., Friederichs, W. J., & März, M. (2023) · Medical Education Online · doi:10.1080/10872981.2023.2220920 · Cited by 122 · cited 4×
“Despite the remarkable performance of these models on standardized tests, they still commit significant errors, and their performance is often on par with that of novice learners [32,35,36].”
BackgroundInfluentialDas, D., Kumar, N., Longjam, L. A., Sinha, R., Deb Roy, A., Mondal, H., & Gupta, P. (2023) · Cureus · doi:10.7759/cureus.36034 · Cited by 116 · cited 5×
“The development of personalized learning plans and learning materials as well as providing tailored feedback to learners are suggested potential avenues for exploration [1,14,15,20,22,34].”
BackgroundInfluentialPerformance of ChatGPT on the Plastic Surgery Inservice Training Examination
Gupta, R., Herzog, I., Park, J. B., Weisberger, J., Firouzbakht, P., Ocon, V., Chao, J., Lee, E. S., & Mailey, B. A. (2023) · Aesthetic Surgery Journal · doi:10.1093/asj/sjad128 · Cited by 109 · cited 2×
“[26] Lack of up-to-date information “ChatGPT repeats its answers with much confidence and clear explanations, even in case of a totally wrong answer.”
BackgroundChatGPT failed Taiwan’s Family Medicine Board Exam
Weng, T.-L., Wang, Y.-M., Chang, S., Chen, T.-J., & Hwang, S.-J. (2023) · Journal of the Chinese Medical Association · doi:10.1097/jcma.0000000000000946 · Cited by 100 · cited 3×
“[13] Exam preparation/practice Theme 2: Novel learning strategies “The creation of personalized quizzes for students is an illustration of the use of generative AI in medical education evaluations.”
BackgroundChatGPT in Clinical Toxicology
Sabry Abdel-Messih, M., & Kamel Boulos, M. N. (2023) · JMIR Medical Education · doi:10.2196/46876 · Cited by 95 · cited 1×
“However, the range of examples provided in the current literature is relatively limited [1,12,15,31].”
BackgroundGiannos, P. (2023) · BMJ Neurology Open · doi:10.1136/bmjno-2023-000451 · Cited by 87 · cited 1×
“A few studies outline potential scenarios in which educators or learners might interact with AI systems, such as in self-directed learning, simulation environments, and writing assistance [8,11,12,14,19,22,23,25,28,34,35,38,40,64].”
BackgroundChat Generative Pre-trained Transformer: why we should embrace this technology
Chavez, M. R., Butler, T. S., Rekawek, P., Heo, H., & Kinzler, W. L. (2023) · American Journal of Obstetrics and Gynecology · doi:10.1016/j.ajog.2023.03.010 · Cited by 83 · cited 2×
“…text as written by a human, or circumvent traditional learning exercises designed for skill development [12,14,15,19,20,22-25 Many authors emphasize the need for establishing clear-cut policies on the acceptable uses of generative AI within the realm of medical education [14,22,40,42,43].”
BackgroundProof of Concept: Using ChatGPT to Teach Emergency Physicians How to Break Bad News
Webb, J. J. (2023) · Cureus · doi:10.7759/cureus.38755 · Cited by 80 · cited 3×
“This was achieved by simulating patient reactions and dialogues during the disclosure of a new cancer diagnosis [16].”
BackgroundMethodInfluentialGabrielson, A. T., Odisho, A. Y., & Canes, D. (2023) · Journal of Urology · doi:10.1097/ju.0000000000003383 · Cited by 76 · cited 2×
“A significant issue lies in data accuracy, with many authors drawing attention to this technology’s propensity to “hallucinate,” or create false information, and its knowledge being confined to the training data set [1,10,21,22,25,27,28,34,37,44].”
BackgroundMasters, K. (2023) · Medical Teacher · doi:10.1080/0142159x.2023.2208731 · Cited by 76 · cited 4×
“A significant issue lies in data accuracy, with many authors drawing attention to this technology’s propensity to “hallucinate,” or create false information, and its knowledge being confined to the training data set [1,10,21,22,25,27,28,34,37,44].”
BackgroundInfluentialA Philosophical Approach to Addressing Uncertainty in Medical Education
Tonelli, M. R., & Upshur, R. E. G. (2019) · Academic Medicine · doi:10.1097/acm.0000000000002512 · Cited by 73 · cited 1×
“Learning about AI interactions may shed light on how we engage with other artifacts or individuals in the clinical environment, compelling learners to ponder what “accuracy” means in a clinical or learning context [51].”
BackgroundGhosh, A., & Bir, A. (2023) · Cureus · doi:10.7759/cureus.37023 · Cited by 67 · cited 3×
“A significant issue lies in data accuracy, with many authors drawing attention to this technology’s propensity to “hallucinate,” or create false information, and its knowledge being confined to the training data set [1,10,21,22,25,27,28,34,37,44].”
BackgroundAye, AI! ChatGPT passes multiple-choice family medicine exam
Morreel, S., Mathysen, D., & Verhoeven, V. (2023) · Medical Teacher · doi:10.1080/0142159x.2023.2187684 · Cited by 56 · cited 2×
“A significant issue lies in data accuracy, with many authors drawing attention to this technology’s propensity to “hallucinate,” or create false information, and its knowledge being confined to the training data set [1,10,21,22,25,27,28,34,37,44].”
BackgroundJeffery, R., Navarro, T., Lokker, C., Haynes, R. B., Wilczynski, N. L., & Farjou, G. (2012) · Journal of Medical Internet Research · doi:10.2196/jmir.2105 · Cited by 50 · cited 1×
“Medical textbooks, often considered the gold standard in medical knowledge, are perpetually outdated, are limited in scope, and may contain inaccuracies [52-54].”
BackgroundWhat could we make of AI in plastic surgery education
Koljonen, V. (2023) · Journal of Plastic, Reconstructive & Aesthetic Surgery · doi:10.1016/j.bjps.2023.04.055 · Cited by 44 · cited 3×
“AI image generation technology has also been used in 2 distinct studies [17,18].”
BackgroundMethodFinding the Place of ChatGPT in Medical Education
van de Ridder, J. M. M., Shoja, M. M., & Rajput, V. (2023) · Academic Medicine · doi:10.1097/acm.0000000000005254 · Cited by 41 · cited 5×
“The precision and trustworthiness of generative AI are fundamental concerns thoroughly elaborated in many publications [8,11-15, 20,22, 24, 32, The propensity of these systems to generate and propagate misinformation is a notable risk.”
BackgroundInfluentialThe next paradigm shift? ChatGPT, artificial intelligence, and medical education
Wang, L. K.-P., Paidisetty, P. S., & Cano, A. M. (2023) · Medical Teacher · doi:10.1080/0142159x.2023.2198663 · Cited by 36 · cited 1×
“Furthermore, an overemphasis on AI-based learning opportunities could diminish human interaction and engagement, which are fundamental to learning and honing patient-interaction skills [22,47].”
BackgroundArtificial scholarship: LLMs in health professions education research
Ellaway, R. H., & Tolsgaard, M. (2023) · Advances in Health Sciences Education · doi:10.1007/s10459-023-10257-4 · Cited by 36 · cited 10×
“Moreover, concerns have arisen that generative AI may produce biased content or lack representation of all populations [8,11-15, 19, 20, 22, 24, This inability to observe how a response is generated is often referred to as “the black box” problem [49].”
BackgroundInfluentialWill ChatGPT Match to Your Program?
Zumsteg, J. M., & Junn, C. (2023) · American Journal of Physical Medicine & Rehabilitation · doi:10.1097/phm.0000000000002238 · Cited by 35 · cited 6×
“Several articles voice concerns about potential academic dishonesty [12,14,15,19,20,22,23,25,40-43].”
BackgroundInfluentialRethinking assessment in response to generative artificial intelligence
Pearce, J., & Chiavaroli, N. (2023) · Medical Education · doi:10.1111/medu.15092 · Cited by 30 · cited 3×
“This sentiment aligns with the viewpoint of Pearce and Chiavoroli [30] that we must rethink our learner assessment methods in a world where generative AI is increasingly prevalent.”
BackgroundTriola, M. M., Reinstein, I., Marin, M., Gillespie, C., Abramson, S., Grossman, R. I., & Rivera, R. (2023) · Academic Medicine · doi:10.1097/acm.0000000000005202 · Cited by 29
The Case Against Race-Based GFR
Uppal, P., Golden, B. L., Panicker, A., Khan, O. A., & Burday, M. J. (2022) · Delaware Journal of Public Health · doi:10.32481/djph.2022.014 · Cited by 28
Gandhi Periaysamy, A., Satapathy, P., Neyazi, A., & Padhi, B. K. (2023) · Annals of Medicine & Surgery · doi:10.1097/ms9.0000000000000371 · Cited by 27 · cited 8×
“[12] Self-directed learning “However, ChatGPT performed acceptably in negative phrase questions, mutually exclusive questions, and case scenario questions, and it can be a helpful tool for learning and exam preparation.””
BackgroundInfluentialHow Reliable Are Medical Textbooks?
Tez, M., & Yildiz, B. (2017) · Journal of Graduate Medical Education · doi:10.4300/jgme-d-17-00209.1 · Cited by 21
Revolutionizing Neurosurgery with GPT-4: A Leap Forward or Ethical Conundrum?
Li, W., Fu, M., Liu, S., & Yu, H. (2023) · Annals of Biomedical Engineering · doi:10.1007/s10439-023-03240-y · Cited by 19 · cited 1×
“A few studies outline potential scenarios in which educators or learners might interact with AI systems, such as in self-directed learning, simulation environments, and writing assistance [8,11,12,14,19,22,23,25,28,34,35,38,40,64].”
BackgroundHuston, J. C., & Kaminski, N. (2023) · ATS Scholar · doi:10.34197/ats-scholar.2022-0141ps · Cited by 15 · cited 3×
“AI image generation technology has also been used in 2 distinct studies [17,18].”
BackgroundMethodRose, C., Barber, R., Preiksaitis, C., Kim, I., Mishra, N., Kayser, K., Brown, I., & Gisondi, M. (2023) · doi:10.2196/preprints.49314 · Cited by 12
Self-citationRose, C. C., Barber, R., Preiksaitis, C. M., Kim, I., Mishra, N., Kayser, K., Brown, I., & Gisondi, M. A. (2023) · Journal of Medical Internet Research · doi:10.2196/49314 · Cited by 12 · cited 1×
“Lack of representation and bias in medical records data are major concerns, and we are only beginning to recognize biases in technology that has been in use in health care for years [57-59].”
BackgroundSelf-citationResponse to: Aye, AI! ChatGPT passes multiple-choice family medicine exam
Masters, K. (2023) · Medical Teacher · doi:10.1080/0142159x.2023.2190476 · Cited by 5 · cited 1×
“As AI usage becomes increasingly prevalent among learners, there may be a need to adapt assessment methods, given the potential effects on the validity of knowledge evaluations [30,46].”
BackgroundGustavii, B. (2017) · Cambridge University Press eBooks · doi:10.1017/9781316650431.015 · Cited by 1
(2022) · doi:10.32481/djph.2022.08.014
OpenAI, Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D., Altenschmidt, J., Altman, S., et al. (2023) · arXiv · doi:10.48550/arxiv.2303.08774
Research works printed but not linked: 3
- 10.6145/jme.202303_27(1).0005 · doi:10.6145/jme.202303_27(1).0005
- OpenAIGPT-4 Technical ReportarXiv202303272023-06-01https://arxiv.org/abs/2303.08774v3
- Bloom, BS, 1956, Taxonomy of Educational Objectives: The Classification of Educational Goals
Cited by
Generative AI Literacy: Twelve Defining Competencies
Annapureddy, R., Fornaroli, A., & Gatica-Perez, D. (2025) · Digital Government: Research and Practice · doi:10.1145/3685680 · Cited by 333
“Last, incorporating differentiation between generative and predictive models into professional training programs and educational initiatives is crucial to address the challenges of each model type and be better prepared for using generative AI in typical applications, or in specific careers in…”
Moulaei, K., Yadegari, A., Baharestani, M., Farzanbakhsh, S., Sabet, B., & Reza Afrash, M. (2024) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2024.105474 · Cited by 244
Takita, H., Kabata, D., Walston, S. L., Tatekawa, H., Saito, K., Tsujimoto, Y., Miki, Y., & Ueda, D. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-01543-z · Cited by 203
Pedagogical Applications of Generative AI in Higher Education: A Systematic Review of the Field
Qian, Y. (2025) · TechTrends · doi:10.1007/s11528-025-01100-1 · Cited by 206
Aljamaan, F., Temsah, M.-H., Altamimi, I., Al-Eyadhy, A., Jamal, A., Alhasan, K., Mesallam, T. A., Farahat, M., & Malki, K. H. (2024) · JMIR Medical Informatics · doi:10.2196/54345 · Cited by 211
“Additional factors might include insufficient training data for the LLM [17,35,36], context misinterpretation, lack of external validation, and overreliance on pattern recognition, all of which could contribute to variable degrees of hallucination, depending on the specific LLM and prompt structure.”
Background
Metadata
- Authors
- Preiksaitis, C., & Rose, C.
- Affiliations
- Stanford Medicine (US)
- Corresponding author
- Preiksaitis, Carl
- Venue
- JMIR Medical Education
- Publisher
- JMIR Publications Inc.
- Published
- 2023-10-20
- Volume
- 9
- Pages
- e48785
- ISSN
- 2369-3762 (ISSN-L)
- PMC
- PMC10625095
- OpenAlex
- W4387144847
- Cited by
- 552
- References
- 68
- Citations by year
- 2026: 238 · 2025: 204 · 2024: 104 · 2023: 5
- FWCI
- 5.54
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Locations
- JMIR Medical Education (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Social Sciences › Education › Artificial Intelligence in Education
- Social Sciences › Education › Innovations in Medical Education
- Keywords
- generative AI, medical education, ChatGPT, academic integrity, self-directed learning, simulation-based training, writing support, critical evaluation of AI, assessment methods, human-AI interaction, AI competencies
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-15, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-15 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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.