bims-aimedu 2026-03-08 papers

bims-aimedu

Biomed News

on AI in medical education

Issue of 2026–03–08
ten papers selected by
Angela Spencer, Saint Louis University

Public Perceptions of Artificial Intelligence in Medicine and Implications for Future Medical Education: A Cross-Sectional Survey.
Advantages and challenges for utilization of generative artificial intelligence in clinical nursing practice: an integrative review.
Application of AI-Generated Content in Medical Education: Systematic Review of the Impact on Critical Thinking Abilities of Medical Students.
Perception of Medical Speciality Professors towards Including Artificial Intelligence in their Curricula.
Innovating instructional design through generative AI prompt engineering for health professions educators.
Integrating Sustainable Development Goals (SDGs) and Generative AI to Enhance Language Digital Literacy and Creativity in EFL Learning Environments.
How latent and prompting biases in AI-generated historical narratives influence opinions.
Developing professional ethical guidance for healthcare AI use (PEG-AI): an attitudinal survey pilot.
Improving Evidence-Based Practice Through the Development and Validation of Key Wo\rds to Identify Intervention Articles in Communication Disorders.
Improving Clinicians' Digital and Artificial Intelligence-Related Competence Within Healthcare Organizations in the United States: A Strategic Framework.

JMIR Form Res. 2026 Mar 04.

Public Perceptions of Artificial Intelligence in Medicine and Implications for Future Medical Education: A Cross-Sectional Survey.

Michael Kirchberger.

BACKGROUND: The integration of artificial intelligence (AI) into clinical practice is contingent on public trust. This trust often depends on physician oversight, yet a significant gap exists between the need for AI-competent physicians and the current state of medical education. While the perspectives of students and experts on this gap are known, the views of the US general public remain largely unquantified.
OBJECTIVE: This study aimed to assess US public perceptions regarding AI in medicine and the corresponding, emergent needs for medical education. We specifically sought to quantify public trust in different diagnostic scenarios, concerns about physician over-reliance on AI, support for mandatory AI education, and priorities for the future focus of medical training.
METHODS: We conducted a cross-sectional, web-based survey of US adults in November 2025. Participants (N=524) were recruited via SurveyMonkey Audience. We calculated descriptive statistics, frequencies (n), proportions (%), and 95% confidence intervals (CIs) for all main survey items.
RESULTS: A total of 524 participants completed the survey. A majority (62.8%, 329/524; 95% CI 58.6%-66.9%) placed the most trust in a physician's diagnosis based on their expertise alone; only 7.8% (41/524; 95% CI 5.5%-10.1%) trusted an AI-first diagnostic model. Trust was highly contingent on training: 93.9% (492/524) of participants rated formal physician training on AI limitations as "Essential" or "Very important." Widespread concern about physician over-reliance on AI was reported, with 81.1% (425/524) being "Very" or "Extremely concerned." Consequently, 85.2% (446/524) agreed or strongly agreed that training on AI use, ethics, and limitations should be mandatory in medical school. When asked about future educational priorities, 70.2% (368/524; 95% CI 66.3%-74.1%) believed medical education should focus on human-centered skills (eg, empathy, communication) over clinical skills.
CONCLUSIONS: The US public expresses conditional trust in medical AI, strongly preferring physician-led and critically supervised models. These findings reveal a clear public mandate for medical education reform. The public expects future physicians to be mandatorily trained to appraise AI, understand its limitations, and refocus their professional development on the human-centered skills that technology cannot replace.
CLINICALTRIAL:

DOI: https://doi.org/10.2196/89123
BMC Nurs. 2026 Mar 04.

Advantages and challenges for utilization of generative artificial intelligence in clinical nursing practice: an integrative review.

Ga In Han, Hyeyeon Choi, Youn-Jung Son.

   BACKGROUND: Generative artificial intelligence (AI) has the potential to ease the administrative burden placed on nurses and to advance the quality and efficiency of care. Emerging nursing evidence suggests that increasing clinical complexity is associated with missed care, indicating that generative AI may play a role in supporting clinical judgment and streamlining nursing workflows. Despite these possibilities, its use within nursing area remains at an early developmental stage. This study aims to explore current applications of generative AI in clinical nursing, and to identify its advantages and challenges.
METHODS: This integrative review followed the updated methodology of Whittemore and Knafl in 2005. Individual studies in this review were selected based on their relevance to the utilization of generative AI in clinical nursing practice. A literature search was conducted using PubMed, Cochrane, CINAHL, Web of Science, EMBASE, SCOPUS, and Google Scholar, covering the period from January 2000 to September 2025. A quality assessment was performed using a mixed-methods appraisal tool.
RESULTS: The 15 included studies, which were published between 2023 and 2025, comprised randomized controlled trials, cross-sectional studies, qualitative studies, and mixed-methods studies. In clinical nursing practice, generative AI is mainly utilized in three areas: clinical decision-making support, patient education and self-management support for chronic diseases, and efficiency of nursing work. The most common purpose of generative AI is to enhance nursing efficiency. ChatGPT is most frequently used in clinical decision-making support and enhancing nursing workflows, while task-oriented chatbots are primarily applied to patient education and self-management. Generative AI requires enhanced accuracy and reliability through continuous learning from new data, empathic conversations, and human interaction.
CONCLUSIONS: Our findings suggest that the use of generative AI in nursing practice has the potential to support clinical decision-making, educate patients and enable self-management, and improve nursing efficiency. By reducing documentation burdens, optimizing workflows and enabling personalized care, generative AI could enhance nursing practice. However, these findings should be interpreted cautiously given the heterogeneity of study designs and the predominantly exploratory nature of the available evidence. The integration of generative AI into nursing practice requires continued improvements in accuracy, reliability, empathic interaction and meaningful human involvement. These potential benefits can be realized by nurses, with appropriate competencies and a critical understanding of both the advantages and limitations of generative AI being fostered.
CLINICAL TRIAL NUMBER: Not applicable.

Keywords:  Generative artificial intelligence; Hospitals; Nurses; Practical nursing; Review literature as topic

DOI:  https://doi.org/10.1186/s12912-026-04518-x
JMIR Med Educ. 2026 02 27. 12 e79939

Application of AI-Generated Content in Medical Education: Systematic Review of the Impact on Critical Thinking Abilities of Medical Students.

Jinlei Li, Fen Ai, Jueyan Wang, Bingxin Cheng, Yu Li, Zhen Chen.

   Background: With the rapid development of artificial intelligence technology, artificial intelligence-generated content (AIGC) is increasingly widely applied in the field of medical education. Large language models, such as ChatGPT, are a prominent type of AIGC technology. Critical thinking is a core ability in medical education, but the impact of AIGC technology on the critical thinking ability of medical students remains unclear. Medical students are at a crucial stage in cultivating critical thinking, and the intervention of AIGC technology may have a profound impact on this process.
Objective: This study aims to systematically review the impact of AIGC technology on the complex mechanisms affecting medical students' critical thinking abilities and build a corresponding strategic framework. The findings are intended to provide theoretical support and practical guidance for applying AIGC in medical education.
Methods: This study followed 2020 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, with the retrieval scope limited to English studies published between November 2022 and June 2025. Through the PubMed database, combined with the search methods of subject terms and free words, relevant studies involving the impact of AIGC on the critical thinking of medical students were screened for using keywords such as "AIGC," "medical students," and "critical thinking." Two independent reviewers screened and evaluated the literature, and ultimately conducted a qualitative analysis based on the common themes extracted from the literature.
Results: AIGC technology in medical education is 2-fold. First, AIGC's powerful information capabilities provide abundant learning resources and efficient tools. This accelerates knowledge acquisition and broadens learning scope. Second, overreliance on AIGC may lead to mental inertia, weaken critical thinking skills, and cause academic integrity issues among students. Research has found that strategies such as customized AIGC tools, virtual standardized patients, new models of resource integration, and proactive assessment of AI limitations can effectively make up for the deficiencies of AIGC in cultivating high-level critical thinking, helping medical students maintain and enhance their critical thinking and problem-solving abilities.
Conclusions: AIGC technology application in medical education needs to carefully weigh the pros and cons. By optimizing the design and usage of AIGC tools and combining them with the guidance and supervision of educators, they can be transformed into powerful tools for promoting the development of critical thinking among medical students. Future research should further expand the scope of study, optimize research methods, pay attention to individual differences, track long-term effects, and deeply explore the influence of ethical and cultural factors to more comprehensively assess the application potential and challenges of AIGC technology in medical education.

Keywords:  AIGC; ChatGPT; LLM; artificial intelligence–generated content; critical thinking; large language models; medical students

DOI:  https://doi.org/10.2196/79939
Niger Postgrad Med J. 2026 Mar 01. 33(2): 232-238

Perception of Medical Speciality Professors towards Including Artificial Intelligence in their Curricula.

Hid Felizardo Cordero-Franco, Joahannah Mynervah Carrillo-García, Miguel Angel Romero-García, Francisco Javier Guzmán-de la Garza.

   BACKGROUND: Medical residents have shown interest in artificial intelligence (AI) for their clinical and academic activities; however, research on the perception of speciality professors regarding AI is lacking.
AIMS: To estimate the proportion of medical professors who perceive it necessary to include AI in the curricula of their specialities, along with their self-efficacy, perceived benefits and identified barriers to AI tools.
SUBJECTS AND METHODS: Cross-sectional study of 108 medical speciality professors in Monterrey, Mexico, in June 2024. We sent them a self-administered questionnaire that included questions about the professors' perceptions of the need to incorporate AI in the specialities' curricula, their self-efficacy, benefits, barriers to AI and other variables. We estimated relative frequencies, 95% confidence intervals (CI), measures of central tendency and dispersion. Self-efficacy, benefits and barriers scores were compared by interest variables using the Mann-Whitney or Student's t-tests. P < 0.05 was significant.
RESULTS: Ninety-eight professors (90.7%, 95% CI: 85.3-96.2) considered it necessary to incorporate AI into their academic programmes. Self-efficacy scores were higher among those working in the private sector and those with a course, workshop or diploma in AI. In addition, those with a master's degree or a doctorate, who had completed a course, workshop or diploma, showed higher benefits scores.
CONCLUSIONS: Nine out of ten professors of medical specialities believe it is necessary to include AI in their curricula, with a high perception of its self-efficacy and benefits. Academic authorities in medical specialisation should examine the growing integration of AI in modern educational environments.

Keywords:  Artificial intelligence; curriculum; faculty; internship and residency; self-efficacy; specialisation

DOI:  https://doi.org/10.4103/npmj.npmj_389_25
Acad Med. 2025 Dec 06. pii: wvaf082. [Epub ahead of print]

Innovating instructional design through generative AI prompt engineering for health professions educators.

Taralyn Tan, Krisztina Fischer.

   PROBLEM: Teaser: An experiential learning intervention to train medical educators to effectively engage generative AI for instructional design is described.Theory-informed and evidence-based educational offerings promote student learning and equity but are time-consuming and require health professions educators to have content expertise in inclusive instructional design. While -generative AI (GAI) offers the potential to overcome these barriers, educators must learn to effectively leverage GAI tools for evidence-based instructional design. In this work, the authors piloted and evaluated a 2-part experiential learning activity to equip educators to effectively engage with GAI for instructional design purposes.
APPROACH: The authors implemented the GAI innovation in the graduate-level "Teaching 100" course (enrollment n = 27) at Harvard Medical School September-November 2023. Educators used GAI to annotate their lesson plans to identify application of, and opportunities to incorporate, evidence-based principles of teaching and learning. The 2-part assignment provided scaffolded instruction on prompt engineering and engaged learners in metacognitive reflection on AI-generated content. The authors evaluated the effectiveness of the GAI innovation according to the Kirkpatrick Model: descriptive analysis of self--reflections evaluated educators' subjective experience (Level 1) and planned behavioral changes (Level 3), while quantification of prompt quality pre-/post-instruction measured educators' learning (Level 2).
OUTCOMES: Among educators who completed the 2-part assignment (n = 17/27, 62% completion rate), the quality of -educator-generated AI prompts improved following instruction in prompt engineering: pre-instruction 1.4 (1.2) (mean [SD]) vs post-instruction 4.0 (0.8). The difference in means (2.6 points) was statistically significant (P < .0001, 95% CI [1.9, 3.3]). Metacognitive reflections revealed specific actions educators planned to pursue to implement GAI feedback to improve their instructional design. Educators reported that AI-based assignments enhanced their learning.
NEXT STEPS: The authors are developing a stand-alone, interactive GAI tool to be broadly deployed as a faculty development instructional design resource. This future work will yield a scalable solution to the challenge of developing AI literacy among health professions educators to leverage GAI for theory-informed and evidence-based instructional design.

Keywords:  faculty development; generative AI; instructional design; metacognition; prompt engineering

DOI:  https://doi.org/10.1093/acamed/wvaf082
J Vis Exp. 2026 Feb 13.

Integrating Sustainable Development Goals (SDGs) and Generative AI to Enhance Language Digital Literacy and Creativity in EFL Learning Environments.

Abdul Khalique Khoso, Wang Honggang, Syeda Rabia Tahir, Aasia Nusrat, Muhammad Younas.

This study investigates the integration of Sustainable Development Goals (SDGs) and generative AI, specifically ChatGPT, to enhance language digital literacy, creativity, and motivation in EFL learning environments. Adopting a quantitative, cross-sectional design, the data for this study were collected from n = 420 undergraduate EFL students at one public sector university in China, using validated scales to measure SDG integration (SDG), Use Generative AI like ChatGPT (UCGPT), digital literacy (DL) EFL Students' creativity (SC), EFL Students' motivation (SM) and language learning engagement (LLE). Partial least squares structural equation modeling (PLS-SEM) was employed to analyze direct, mediating, and moderating effects. Findings revealed that SDG integration and UCGPT were significantly associated with SM, which in turn predicted digital literacy and EFL students' creativity. Moreover, LLE moderated the relationship between DL and SC, while SM mediated the effects of SDG integration and UCGPT on digital literacy. These results highlight the transformative potential of integrating SDGs and AI tools in EFL education, providing scalable strategies to develop 21st-century skills. Additionally, the study provides concrete strategies for EFL teachers, such as tasking students with using ChatGPT to prepare for debates on SDG topics, like climate justice, or to co-create multilingual social media campaigns for sustainability awareness. These applications demonstrate how to operationalize SDG-AI integration in everyday lesson planning.

DOI: https://doi.org/10.3791/69445
PNAS Nexus. 2026 Mar;5(3): pgag022

How latent and prompting biases in AI-generated historical narratives influence opinions.

Matthew Shu, Daniel Karell, Keitaro Okura, Thomas R Davidson.

  Large language models (LLMs) can be used to persuade people on a range of issues, particularly through user-driven strategies such as personalizing messages and dialogues intended to change minds. However, their capacity to influence opinions through subtle, latent ideological framing remains relatively understudied. We investigate whether AI-generated historical summaries affect social and political opinions through a preregistered experiment (N = 1,912). Participants read Wikipedia or GPT-4o summaries of two historical events, with AI summaries maintaining factual accuracy while exhibiting different types of framing biases. Default AI summaries led to more liberal opinions compared with Wikipedia, demonstrating the persuasive capability of LLM's latent biases. Summaries purposefully induced with a liberal framing also led to more liberal opinions, regardless of readers' ideologies. Summaries constructed with a conservative framing produced conservative shifts primarily among conservative readers. These findings demonstrate that the use of AI for learning history can influence opinions through both intrinsic and intentional framing mechanisms, even when the content remains factually accurate. As AI becomes integral to information acquisition, recognizing pathways of influence based not only on user-manipulated content but also on models' latent biases is essential for understanding AI's broader societal impacts.

Keywords:  AI framing; AI history; AI persuasion; generative artificial intelligence; large language models

DOI:  https://doi.org/10.1093/pnasnexus/pgag022
AI Soc. 2025 Oct;40(7): 5439-5455

Developing professional ethical guidance for healthcare AI use (PEG-AI): an attitudinal survey pilot.

Helen Smith, Jonathan Ives.

  Healthcare professionals currently lack guidance for their use of AI. This means they currently lack clear counsel to aid their navigation of the problematic novel issues that will arise from their use of these systems. This pilot study gathered and analysed cross-sectional attitudinal and qualitative data to address the question: what should be in professional ethical guidance (PEG) to support healthcare practitioners in their use of AI? Our survey asked respondents (n = 42) to review 6 themes and 15 items of guidance content for our proposed PEG-AI. The attitudinal data are presented as simple numerical analysis and the accompanying qualitative data were subjected to conventional content analysis; the findings of which are presented in this report. The study data allowed us to identify further items that could be added to the PEG-AI and to test the survey instrument for content and face validity prior to wider deployment. Subject to further funding, we plan to take this work further to a wider study involving the next iteration of this survey, interviews with interested parties regarding PEG-AI, and an iterative Delphi process (comprising an initial co-creation workshop followed by iterative consensus building) to enable experts to reach consensus regarding recommendations for the content of PEG for AI use in healthcare. We aim for this work to inform the healthcare regulators as they develop regulatory strategies in this area.

Keywords:  Artificial Intelligence; Ethics; Healthcare; Professional guidance; Qualitative research

DOI:  https://doi.org/10.1007/s00146-025-02276-z
Lang Speech Hear Serv Sch. 2026 Mar 06. 1-16

Improving Evidence-Based Practice Through the Development and Validation of Key Wo\rds to Identify Intervention Articles in Communication Disorders.

Sonia Islam Nisha, Qingyun Wang, Charles Yu, Heng Ji, Alison Eisel Hendricks.

PURPOSE: Although evidence-based practice (EBP) promotes better clinical practice, implementing it in speech-language pathology is challenging. A limited number of intervention studies and inconsistent use of key words in abstracts complicates the implementation of EBP. Given time constraints in clinical practice, identifying relevant research from study abstracts would provide an efficient method for identifying intervention studies. In this project, we compare the classification accuracy of an artificial intelligence (AI)-created set of key words and a set of evidence-based key words that were developed through an analysis of commonly used words in intervention abstracts from American Speech-Language-Hearing Association (ASHA) journals.
METHOD: Abstracts from three major ASHA journals were crawled using Selenium and WebDriver Manager, creating a large database of abstracts for analysis. In Study 1, a random sample of 180 abstracts was annotated as reporting on intervention or nonintervention studies to develop a set of evidence-based key words. In Study 2, classification accuracy was calculated to validate these key words by comparing them with a set of AI-generated (ChatGPT-4.0) key words.
RESULTS: The results suggested that 12%-15% of studies report on an intervention study. Evidence-based key words had higher sensitivity (85%) and a higher positive predictive value (27%) but a lower specificity rate (67%). Also, the likelihood ratio suggested that evidence-based key words had a moderate capacity to identify nonintervention studies (true negatives) accurately.
CONCLUSIONS: Evidence-based key words have the potential to accurately classify intervention studies, addressing a key barrier to EBP implementation in clinical practice. Future research should focus on refining key words by integrating AI and promoting standardized journal reporting by incorporating uniform key words in abstracts.

DOI: https://doi.org/10.1044/2025_LSHSS-25-00144
Appl Clin Inform. 2026 Mar 06.

Improving Clinicians' Digital and Artificial Intelligence-Related Competence Within Healthcare Organizations in the United States: A Strategic Framework.

Milin Patel Patel, Thomas Scharfenberger, Jonathan Mashieh, Shitij Arora, Sunit Jariwala.

INTRODUCTION: In recent years, there has been an emerging wave of artificial intelligence (AI) and digital tools in healthcare, thereby revolutionizing clinical practice. As health systems are increasingly utilizing these tools as a means to improve clinical and operational performance outcomes, it becomes imperative to train and professionally develop key frontline stakeholders, such as clinicians, in digital health and clinical AI to ensure seamless and responsible adoption within healthcare settings.
METHODS: This paper presents a multi-level framework for healthcare systems to effectively integrate digital and AI tools by enhancing clinicians' proficiency and confidence in utilizing these emerging technologies. Framework Components: Our strategic framework consists of three integrated elements: (1) structured training programs to enhance clinicians' understanding and utilization of digital and AI tools; (2) leadership development pathways to cultivate champions who can drive implementation within clinical departments; and (3) performance management processes to ensure sustainable adoption aligned with organizational goals. This multi-level framework addresses current gaps in clinician preparedness for the digital health ecosystem.
CONCLUSIONS: The integration of digital and AI technologies into clinical practice requires systematic approaches to clinician development. By implementing multi-level training, fostering digital-based leadership, and establishing appropriate performance evaluation metrics, healthcare systems can better prepare their workforce for responsible technology adoption. This paper provides actionable strategies that healthcare organizations can adapt to their specific contexts to maximize the potential benefits of digital innovations in patient care.

DOI: https://doi.org/10.1055/a-2828-0479