Current Landscape of Curriculum Development and Implementation in Medical Artificial Intelligence: A Scoping Review Authors: Yue Wang, He Wang, Ting Wang, Xiaofei Ye, Ruixue Mao Published: 2026 ABSTRACT PurposeTo explore the current status of medical AI educational programs and to review how these programs were developed, implemented, and improved.MethodsFollowing the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines, this study searched English articles from 2021 to April 2026 published in PubMed Central, Web of Science, Scopus, and MEDLINE databases. All curricula included in this review were implemented programs, not theoretical frameworks.ResultsAmong the 7652 screened documents, 36 were identified. Data extraction focused on programs’ status, construction methods, and course frameworks. The number of programs has increased rapidly but remains small, with most in the pilot stage. Learners were predominantly medical students, with small class sizes. Curriculum development mainly relied on expert experience. Learning objectives mainly focused on the understanding level, but also covered advanced cognitive skills, and the content, covering AI principles, clinical applications, and ethics, demanded high cognitive engagement. Most programs were offered as modular elective courses in online or face-to-face formats, employing diverse teaching methods and activities. The learning outcomes and course evaluations mainly relied on student feedback. This review identified persistent gaps: the lack of standardized curriculum frameworks, affective learning objectives, and objective assessment tools; limited content on AI development, applications, collaboration, communication; weak integration with existing curricula; and insufficient involvement of learning designers.ConclusionMedical AI education programs are still in the initial stage. There is a need to accelerate curriculum development pathways, enhance practical skill trainings and affective competency developments, and establish effective instructional design and evaluation systems. Introduction The rapid growth and widespread adoption of artificial intelligence (AI) are transforming how healthcare is delivered. As AI technologies are increasingly integrated into clinical specialties such as radiology, pathology, and dermatology,1 there is an emerging demand for AI-related competencies among healthcare professionals. Clinicians need to understand the core principles of AI in medicine and gain the knowledge and skills to critically appraise and ethically incorporate AI tools into their practice.2 However, existing research suggests that the medical industry currently faces an AI talent gap, underscoring the urgent need to cultivate professionals capable of developing, implementing, managing, and evolving AI medical technologies to fully realize AI’s potential.3 Medical AI education offers a critical pathway for building AI literacy among healthcare trainees.4 As a result, the design, development, and implementation of AI-focused educational programs have become key priorities in medical education. In recent years, many organizations have issued policies or guidelines for medical AI education, such as the Principles for the Responsible Use of Artificial Intelligence in and for Medical Education5 proposed by the American College of Physicians, Artificial Intelligence in Health Professions Education assessment6 released by the Association for Medical Education in Europe, and the requirements and guidance for medical education in the EU Artificial Intelligence Act (2024).7 These policies or guidelines highlight the importance of medical AI education and provide broad strategic direction, but they lack specific guidance for designing and developing medical AI courses. At the moment, healthcare organizations advocate for AI-driven medical training,8,9 and many medical schools are actively developing or refining AI curricula.10 Medical professionals, educators, and students also hold positive attitudes toward medical AI courses.11 Despite the great enthusiasm for AI medical education, the lack of a standardized framework for medical AI education,12 as well as educators’ insufficient understanding and confidence in AI,13 hinders the integration of AI content with medical education. Miller’s Pyramid and other models provide a theoretical basis for medical education institutions to develop courses. Institutions can divide AI capabilities into different levels and integrate them more effectively into the existing curriculum system. For example, beginners need to enhance their AI literacy and understand and master AI-related conceptual knowledge; competent users need to possess AI capabilities, including the ability to apply AI systems in medical environments and improve medical treatment quality; experts need to possess greater professional skills, including the development and systematic improvement of AI systems, thereby leading medical AI innovation.14 Additionally, some studies have explored how to integrate AI content into undergraduate, graduate, and continuing medical education.15 For instance, Crotty et al16 suggested a modular approach to teaching, while Ma et al17 advocated conducting basic, practical, experimental, and ethical teaching across the preclinical, clinical, and clinical research stages. These efforts provide references for integrating AI into existing medical education. However, the heterogeneity of medical education systems in different countries18 further affects the design of medical AI courses. Differences in academic duration, governance methods, and clinical practice will all influence AI learning objectives, teaching methods, and integration approaches. For medical schools, key challenges include the contextual development of interdisciplinary AI curricula, the improvement of instructional resources, and the enhancement of AI competencies among healthcare professionals. Although previous studies have explored medical AI education, most have focused on the theoretical construction of courses or on improving medical students’ knowledge, skills, and attitudes, while rarely addressing programs design and practical processes. Moreover, the course samples selected for research on program design and practical processes were limited. They often included theoretical frameworks that had not been implemented, thus failing to reflect the actual programs situation. Therefore, in order to comprehensively understand the current development and implementation status and problems of medical AI education programs, this study chose the scoping review method, widely including literature targeting different learning objects and different training levels,19 based on one of the most recognized frameworks for designing, implementing, and evaluating courses in the medical education field, Kern’s Six-Step Approach to Curriculum Development,20 and conducted a structured analysis of the included literature, with a focus on the key areas consistent with this framework: learning objectives, scope and organization of teaching content, teaching implementation (teaching environment, teaching strategies), learning outcome assessment method, and curriculum evaluation and feedback. This study aims to review the current status of medical AI education programs’ construction and provide corresponding support and guidance for medical schools to design and develop AI courses. Materials and Methods Study Design This study was conducted based on the five-stage process proposed by Levac et al21 The steps in the review process included defining the research questions, identifying the relevant literature, selecting the eligible studies, extracting and charting the data, and synthesizing the results. Results were reported in accordance with the PRISMA-ScR. The research questions guiding this study primarily include: (1) how existing medical AI courses and training programs are developed, (2) which learner populations they target, (3) what learning objectives are commonly defined (4) what content is included and how curriculum contents are structured, (5) how programs are delivered—including teaching environment, teaching strategies, and learning outcome assessment methods, (6) and what approaches are used to evaluate and improve the course. Literature Search The search was conducted between March and April 2026. The search strategy used terms such as “artificial intelligence”, “machine learning”, “deep learning”, “medical curriculum”, “medical competency”, and “medical literacy”. Publications were retrieved from 2021 to April 2026 in four bibliographic databases: PubMed Central, Web of Science, Scopus, and MEDLINE. Study Selection All the retrieved references were first imported into Rayyan for duplicate removal. Two reviewers then screened titles and abstracts preliminarily through Rayyan. References with consistent screening results were retained, and discrepancies were resolved by a third reviewer. After the initial screening, the two reviewers evaluated the full texts of the references based on the research question and finally determined which references would undergo systematic analysis. Eligibility Criteria The search only included English journal articles available in the aforementioned databases. The articles must describe structured medical AI programs for medical professionals, including undergraduate, graduate, postgraduate, and continuing medical education. Pre-medical programs were excluded. The medical AI education programs must have been implemented in practice; those developed solely through theoretical methods—such as the Delphi technique, literature review, survey—without actual delivery were excluded. The articles must at least cover all or part of the learning objectives, content selection, teaching implementation, or evaluation feedback. Those without detailed descriptions were excluded. Quality Appraisal of Individual Studies We did not conduct a formal quality assessment or risk-of-bias analysis for the included studies. The scoping review method aims to summarize and aggregate various types of studies, including qualitative, quantitative, or mixed-methods studies, whether published or unpublished, to explore the breadth of existing research areas and identify gaps, providing a basis for future research directions and prioritization. Usually, quality or risk of bias is not used as an exclusion criterion.22 This review focuses on the current process of developing and implementing medical AI courses rather than evaluating intervention effects. Therefore, not conducting a quality assessment is the methodological choice of this review. Data Extraction For the 36 studies that met the inclusion criteria, two reviewers analyzed and coded the literature based on Kern’s Six-Step Approach to Curriculum Development and the framework proposed by Obadeji.23 They conducted a systematic analysis focusing on the learning objectives of the courses, the course content, the organization of the content, teaching strategies, the implementation methods of the courses (teaching environment, teaching strategies, and evaluation methods of learning outcomes), as well as course assessment and improvement, to ensure the reliability and accuracy of the results, and disagreements were judged by a third reviewer. Results A total of 7652 articles were found. After duplicate removal using Rayyan software and manual verification, 2336 duplicate articles were removed, leaving 5316 articles eligible for title and abstract screening. A total of 5168 irrelevant articles were excluded after title and abstract review, with 148 articles retained for full-text assessment. Sixteen articles were excluded due to inaccessibility, 57 for lacking relevance to curriculum development, and another 39 for designing curricula without practical teaching implementation. Ultimately, 36 studies met all eligibility criteria for final analysis (Figure 1). Figure 1PRISMA-ScR flowchart.The flowchart outlines the process of identifying and screening studies via databases. Initially, records are identified from PubMed Central, Web of Science, Scopus and MEDLINE, with categories such as AI and Medical Curriculum, AI and Medical Literacy, AI and Medical Competency, ML and Medical Curriculum, ML and Medical Literacy, ML and Medical Competency, DL and Medical Curriculum, DL and Medical Literacy and DL and Medical Competency. A total of 2336 duplicate records are removed before screening, leaving 5316 records for screening. Of these, 5168 records are excluded. Next, 148 records are sought for retrieval, with 16 not retrieved. Records assessed for eligibility total 132, with exclusions for no mention of curriculum development (57) and no practical implementation (39). Finally, 36 studies are included in the review.A flowchart of study identification and screening process for a review. PRISMA-ScR flowchart. This scoping review found that all 36 included studies covered the four core elements of Kern’s Six-Step Approach to Curriculum Development (Table 1): learning objectives, teaching content, teaching implementation, learning outcome assessment, and curriculum evaluation. Although the depth and quality of the descriptions in the studies varied to some extent, this result clearly indicates that while the studies did not explicitly adopt Kern’s Six-Step Approach to Curriculum Development framework to describe their curricula, the framework itself has been widely applied, even though such coverage was often superficial rather than substantive.Table 1Analysis of Medical AI Education ProgramsReferenceCountryYearYear of First DeliveryEnrollmentCurriculum DevelopmentTarget AudiencesSubjectsLearning ObjectivesTeaching ContentContent OrganizationTeaching EnvironmentTeaching StrategiesLearning Outcome AssessmentCourse Evaluation and ImprovementZand J et alThe United States2026N/AN/ALiterature review, questionnaire surveysAI specialistsContinuing medical educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentN/AModularN/ALess than one weekOnline asynchronous learningN/ASelf-directed learningCase, hands-on practice, demonstrationIndividual learningFormativeSelf-assessmentObjective evaluationLearner feedbackN/AImplementing formal assessmentMeade SM et alThe United States20232022<60Literature review, questionnaire surveysN/ASchool education, postgraduate medical educationNeurosurgeryKnowledge, skillTechnical skill, ethical, legalCourseIndependentOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learningN/ASelf-directed learningHands-on practice, project-based learningIndividual learning, group learningFormative and summativeTeacher evaluationBehavior evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationStrengthening instructional interactionKrive J et alThe United States20232019<60N/AInterdisciplinary teamsSchool educationN/AKnowledge, skillClinical application, communicationCourseIndependentOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learning, real-time online activitiesN/ASelf-directed learningCase, project-based learningIndividual learning, group learningFormativeTeacher evaluationObjective evaluation, behavior evaluationLearner feedbackOverall course quality and learner satisfactionIntegrating AI into different learning phasesJagannath P et alThe United States20252022>150Questionnaire surveysClinical expertsSchool education, postgraduate medical education, continuing medical educationN/AKnowledgeTechnical skill, clinical applicationLectureIndependentOptionalModularN/AN/AOnline asynchronous learning, real-time online activitiesN/ASelf-directed learningdiscussion, demonstrationCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsReading Turchioe M et alThe United States20242023<60Literature reviewClinical expertsSchool educationNursingKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedCompulsoryModularN/ALess than one weekFace-to-faceProfessional background of instructorsDidactic teachingCase, hands-on practiceCollective learning, individual learningSummativeSelf-assessmentSubjective report, behavior evaluationN/AN/AN/AKhalil AI et alSaudi Arabia20252024>150Literature reviewClinical expertsContinuing medical educationNursingKnowledge, skillTechnical skill, clinical application, ethical, legal, communicationCourseIndependentOptionalModularN/AMore than one monthOnline and face-to-faceN/ADidactic teachingDiscussionCollective learning, group learningSummativeSelf-assessmentSubjective report, behavior evaluationN/AN/AN/AHu R et alCanada20232022<60N/AInterdisciplinary teamsPostgraduate medical educationRadiologyKnowledge, skillTechnical skill, clinical applicationWorkshopIntegratedOptionalModularN/AOne week to one monthOnline and face-to-faceProfessional background of instructorsFlipped classroomCase, demonstrationCollective learningSummativeSelf-assessmentSubjective reportN/AN/AN/ADavidson AE et alThe United States2025202490-150N/AInterdisciplinary teamsN/AN/AKnowledge, skillTechnical skill, clinical application, ethical, legalWorkshopIndependentOptionalModularN/ALess than one weekOnline and face-to-faceProfessional background of instructorsSelf-directed learningHands-on practiceCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackCourse content, and curriculum organizationStrengthening instructional interactionOftring ZS et alGermany2025N/A<60N/AInterdisciplinary teamsSchool educationN/AKnowledgeClinical applicationWorkshopIntegratedOptionalModularClinical clerkshipLess than one weekOnline and face-to-faceN/AFlipped classroomDiscussion, case, demonstration, project-based learning, role-playingCollective learning, individual learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum OrganizationN/AStogiannos N et alThe United Kingdom2025N/AN/AN/AInterdisciplinary teamsSchool educationRadiologyKnowledgeClinical application, ethical, legalSeminarIntegratedOptionalModularClinical clerkshipN/AFace-to-faceN/ADidactic teachingDiscussion, case, demonstrationCollective learningSummativeSelf-assessmentObjective evaluationN/AN/AEnhancing practical components, strengthening instructional interactionTill H et alAustralia20262024<60N/AClinical expertsContinuing medical educationPediatricsKnowledge, skillTechnical skill, clinical applicationWorkshopIndependentOptionalModularN/AN/AFace-to-faceProfessional background of instructorsDidactic teachingN/ACollective learningSummativeSelf-assessmentSubjective reportN/AN/AN/AOta Y et alJapan2026202590-150N/AN/ASchool educationN/AKnowledgeEthical, legal, communicationCourse, workshopIntegratedCompulsoryModularClinical clerkshipLess than one weekFace-to-faceN/ADidactic teaching, self-directed learningProject-based learningCollective learning, group learningFormative and summativeSelf-assessmentObjective evaluation, behavior evaluationLearner feedbackN/AN/ARezazadeh H et alIran20252023N/ALiterature review, questionnaire surveysInterdisciplinary teamsSchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentOptionalN/AN/AN/AOnline and face-to-faceN/AFlipped classroomDiscussion, project-based learning, peer teaching, game-based learningGroup learningSummativeSelf-assessmentSubjective reportLearner feedbackCourse content, and curriculum organizationEnhancing practical componentsElhoshy H et alEgypt20252022>150Literature reviewClinical experts, AI specialists, or authoritative organizationsSchool educationN/AKnowledge, skillClinical applicationCourseIndependentOptionalModularPreclinical theoreticalOne week to one monthOnline asynchronous learning, real-time online activitiesProfessional background of instructorsSelf-directed learningDiscussionCollective learning, individual learning, group learningFormative and summativeSelf-assessmentObjective evaluationLearner and instructor feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsChang L et alThe United States20262024N/AN/AN/ASchool educationN/AKnowledge, skillClinical application, ethical, legal, communication, collaboration, healthcare quality improvementCourseIndependent, IntegratedOptionalModular, spiralPreclinical theoretical, clinical clerkshipN/AOnline asynchronous learningN/AN/ADiscussion, case, hands-on practiceN/AN/ATeacher evaluationSubjective report, behavior evaluationN/AN/AN/ALindqwister et alThe United States2021N/A<60N/AClinical experts, AI specialists, or authoritative organizationsPostgraduate medical educationN/AKnowledgeN/ACourseIndependentCompulsoryModularN/AMore than one monthFace-to-faceProfessional background of instructorsDidactic teaching, self-directed learningDemonstrationCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, strengthening instructional interactionJalal AHB et alThe United Kingdom20242022<60N/AClinical expertsSchool education, continuing medical educationNeurosurgeryKnowledge, skillTechnical skillCourseIndependentOptionalModularN/ALess than one weekOnline asynchronous learning, real-time online activitiesN/AFlipped classroomDiscussion, demonstrationIndividual learning, group learningSummativeSelf-assessmentSubjective report, objective evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationPark SH et alThe United States20242022<60N/AClinical expertsSchool educationN/AKnowledge, skillCourseIntegratedOptionalModularPreclinical theoreticalN/AOnline and face-to-faceN/AFlipped classroomDiscussion, case, demonstrationCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AMihalas GI et alRomania2025N/A<60Questionnaire surveysClinical expertsSchool educationN/AKnowledgeClinical application, ethical, legalCourseIndependentN/AN/AN/AN/AN/AN/ADidactic teachingHands-on practiceCollective learningN/ATeacher evaluationObjective evaluationLearner and instructor feedback, other stakeholders’ feedbackCourse content, and curriculum organizationN/AAbid A et alThe United States20242021<60N/AN/ASchool educationN/AKnowledgeTechnical skill, clinical application, ethical, legalCourseIntegratedOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learningProfessional background of instructorsSelf-directed learningProject-based learningIndividual earningSummativeSelf-assessmentSubjective report, behavior evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/APizzolla I et alBelgium20232022–2023 academic year<60N/ASchool educationN/AKnowledge, skillTechnical skill, clinical applicationCourseIndependentCompulsoryModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningDiscussion, case, hands-on practiceIndividual learningFormativeteacher EvaluationObjective evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, integrating AI into different learning phasesLindqwister et alThe United States2023N/A<60N/AClinical experts, AI specialists, or authoritative organizationsPostgraduate medical educationN/AKnowledgeN/AN/AIndependentOptionalModularN/AN/AOnline and face-to-faceProfessional background of instructorsDidactic teachingDiscussion, project-based learningCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedback, other stakeholders’ feedbackOverall course quality and learner satisfactionN/AHedderich DM et alGermany2021N/A90-150N/AN/ASchool education, postgraduate medical education, continuing medical educationN/AKnowledge, skillClinical application, ethical, legalCourseIndependentOptionalModularN/AMore than one monthOnline and face-to-faceN/AFlipped classroomDiscussion, demonstrationCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AAgarwal G et alThe United States2025N/A>150Literature reviewInterdisciplinary teamsSchool educationNursing, physical therapyKnowledgeClinical application, ethical, legalLectureIndependentN/AModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningCaseIndividual learningFormativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationStrengthening instructional interactionWang A et alNew Zealand2026N/AN/AN/Ainterdisciplinary teamsSchool educationRadiologyKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedCompulsoryModularPreclinical theoreticalN/AN/AProfessional background of instructorsDidactic teachingCaseCollective learningSummativeTeacher evaluationObjective evaluationLearner and instructor feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, implementing formal assessmentTaskiran N et alTurkey20232022>150N/AN/ASchool educationNursingKnowledgeClinical application, ethical, legalCourseIndependentOptionalN/APreclinical theoreticalMore than one monthFace-to-faceN/ADidactic teachingN/ACollective learningSummativeSelf-assessmentSubjective reportN/AOverall course quality and learner satisfactionN/ATejani AS et alThe United States20242021<60Questionnaire surveysClinical expertsPostgraduate medical educationRadiologyKnowledge, skillTechnical skill, clinical application, ethical, legalLectureIntegratedOptionalModularN/AMore than one monthFace-to-faceN/ADidactic teaching, flipped classroomDiscussion, case, hands-on practice, project-based learningCollective learning, group learningN/AN/AN/ALearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AVan Kooten MJ et alThe Netherlands2024N/A<60N/AInterdisciplinary teamsPostgraduate medical educationRadiologyKnowledgeClinical application, ethical, legalCourseIndependentCompulsoryModularN/ALess than one weekN/AProfessional background of instructorsDidactic teachingDiscussion hands-on practiceCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsLaupichler MC et alGermany20222021<60N/AClinical expertsSchool educationN/AKnowledgeClinical applicationCourseIndependentOptionalModularClinical clerkshipMore than one monthOnline asynchronous learning, real-time online activitiesProfessional background of instructorsFlipped classroomDiscussion, hands-on practiceCollective learning, individual learningFormative and summativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/ALevingston H et alThe United States2024N/AN/AN/AN/ASchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentN/AModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningN/AIndividual learningFormativeSelf-assessmentObjective evaluationN/AOverall course quality and learner SatisfactionN/ASchaaf J et alGermany20252025N/AN/AInterdisciplinary teamsSchool education, continuing medical educationN/AKnowledgeClinical application, ethical, legalCourseIndependentN/AModularN/ALess than one weekOnline asynchronous learningN/ASelf-directed learningCase, storytellingIndividual learningFormative and summativeSelf-assessmentObjective evaluationN/AN/AN/AYangöz ŞT et alTurkey2026202490-150N/AN/ASchool educationNursingKnowledge, skillClinical application, ethical, legalCourseIndependentCompulsoryModularN/AMore than one monthFace-to-faceProfessional background of instructorsDidactic teachingDiscussion, hands-on practice, demonstrationN/AFormative and summativeTeacher evaluationObjective evaluationLearner feedbackN/AIntegrating AI into different learning phasesDonkor A et alGhana2025N/A90-150N/AInterdisciplinary teamsSchool educationMedical imagingKnowledgeTechnical skill, clinical application, ethical, legalLectureIndependentOptionalModularPreclinical theoretical, clinical Clerkshipless than one weekFace-to-faceN/ADidactic teachingDiscussionCollective learning, group learningFormativeSelf-assessmentSubjective reportN/AN/AN/AHu R et alCanada20222019>150N/AInterdisciplinary teamsSchool education, postgraduate medical educationN/AKnowledge, skillTechnical skill, clinical applicationWorkshopIndependentOptionalModularPreclinical theoreticalmore than one monthOnline and face-to-faceN/ADidactic teachingDiscussion, case, hands-on practiceCollective learningN/AN/AN/ALearner and instructor feedback, other stakeholders’ feedbackN/AEnhancing practical componentsAlZaabi A et alOman2025N/AN/AN/AN/ASchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedN/ASpiralPreclinical theoreticalN/AN/AN/AFlipped classroomDiscussion, project-based learningGroup learningSummativeSelf-assessmentSubjective reportN/AN/AN/AAbuadas M et alJordan20252024<60N/AInterdisciplinary teamsSchool educationNursingKnowledgeClinical application, ethical, legalCourseIndependentOptionalModularClinical clerkshipLess than one weekN/AN/AN/ADiscussion, case, project-based learningCollective learning, group learningSummativeSelf-assessmentSubjective reportN/AN/AN/A Table 1Analysis of Medical AI Education ProgramsReferenceCountryYearYear of First DeliveryEnrollmentCurriculum DevelopmentTarget AudiencesSubjectsLearning ObjectivesTeaching ContentContent OrganizationTeaching EnvironmentTeaching StrategiesLearning Outcome AssessmentCourse Evaluation and ImprovementZand J et alThe United States2026N/AN/ALiterature review, questionnaire surveysAI specialistsContinuing medical educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentN/AModularN/ALess than one weekOnline asynchronous learningN/ASelf-directed learningCase, hands-on practice, demonstrationIndividual learningFormativeSelf-assessmentObjective evaluationLearner feedbackN/AImplementing formal assessmentMeade SM et alThe United States20232022<60Literature review, questionnaire surveysN/ASchool education, postgraduate medical educationNeurosurgeryKnowledge, skillTechnical skill, ethical, legalCourseIndependentOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learningN/ASelf-directed learningHands-on practice, project-based learningIndividual learning, group learningFormative and summativeTeacher evaluationBehavior evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationStrengthening instructional interactionKrive J et alThe United States20232019<60N/AInterdisciplinary teamsSchool educationN/AKnowledge, skillClinical application, communicationCourseIndependentOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learning, real-time online activitiesN/ASelf-directed learningCase, project-based learningIndividual learning, group learningFormativeTeacher evaluationObjective evaluation, behavior evaluationLearner feedbackOverall course quality and learner satisfactionIntegrating AI into different learning phasesJagannath P et alThe United States20252022>150Questionnaire surveysClinical expertsSchool education, postgraduate medical education, continuing medical educationN/AKnowledgeTechnical skill, clinical applicationLectureIndependentOptionalModularN/AN/AOnline asynchronous learning, real-time online activitiesN/ASelf-directed learningdiscussion, demonstrationCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsReading Turchioe M et alThe United States20242023<60Literature reviewClinical expertsSchool educationNursingKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedCompulsoryModularN/ALess than one weekFace-to-faceProfessional background of instructorsDidactic teachingCase, hands-on practiceCollective learning, individual learningSummativeSelf-assessmentSubjective report, behavior evaluationN/AN/AN/AKhalil AI et alSaudi Arabia20252024>150Literature reviewClinical expertsContinuing medical educationNursingKnowledge, skillTechnical skill, clinical application, ethical, legal, communicationCourseIndependentOptionalModularN/AMore than one monthOnline and face-to-faceN/ADidactic teachingDiscussionCollective learning, group learningSummativeSelf-assessmentSubjective report, behavior evaluationN/AN/AN/AHu R et alCanada20232022<60N/AInterdisciplinary teamsPostgraduate medical educationRadiologyKnowledge, skillTechnical skill, clinical applicationWorkshopIntegratedOptionalModularN/AOne week to one monthOnline and face-to-faceProfessional background of instructorsFlipped classroomCase, demonstrationCollective learningSummativeSelf-assessmentSubjective reportN/AN/AN/ADavidson AE et alThe United States2025202490-150N/AInterdisciplinary teamsN/AN/AKnowledge, skillTechnical skill, clinical application, ethical, legalWorkshopIndependentOptionalModularN/ALess than one weekOnline and face-to-faceProfessional background of instructorsSelf-directed learningHands-on practiceCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackCourse content, and curriculum organizationStrengthening instructional interactionOftring ZS et alGermany2025N/A<60N/AInterdisciplinary teamsSchool educationN/AKnowledgeClinical applicationWorkshopIntegratedOptionalModularClinical clerkshipLess than one weekOnline and face-to-faceN/AFlipped classroomDiscussion, case, demonstration, project-based learning, role-playingCollective learning, individual learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum OrganizationN/AStogiannos N et alThe United Kingdom2025N/AN/AN/AInterdisciplinary teamsSchool educationRadiologyKnowledgeClinical application, ethical, legalSeminarIntegratedOptionalModularClinical clerkshipN/AFace-to-faceN/ADidactic teachingDiscussion, case, demonstrationCollective learningSummativeSelf-assessmentObjective evaluationN/AN/AEnhancing practical components, strengthening instructional interactionTill H et alAustralia20262024<60N/AClinical expertsContinuing medical educationPediatricsKnowledge, skillTechnical skill, clinical applicationWorkshopIndependentOptionalModularN/AN/AFace-to-faceProfessional background of instructorsDidactic teachingN/ACollective learningSummativeSelf-assessmentSubjective reportN/AN/AN/AOta Y et alJapan2026202590-150N/AN/ASchool educationN/AKnowledgeEthical, legal, communicationCourse, workshopIntegratedCompulsoryModularClinical clerkshipLess than one weekFace-to-faceN/ADidactic teaching, self-directed learningProject-based learningCollective learning, group learningFormative and summativeSelf-assessmentObjective evaluation, behavior evaluationLearner feedbackN/AN/ARezazadeh H et alIran20252023N/ALiterature review, questionnaire surveysInterdisciplinary teamsSchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentOptionalN/AN/AN/AOnline and face-to-faceN/AFlipped classroomDiscussion, project-based learning, peer teaching, game-based learningGroup learningSummativeSelf-assessmentSubjective reportLearner feedbackCourse content, and curriculum organizationEnhancing practical componentsElhoshy H et alEgypt20252022>150Literature reviewClinical experts, AI specialists, or authoritative organizationsSchool educationN/AKnowledge, skillClinical applicationCourseIndependentOptionalModularPreclinical theoreticalOne week to one monthOnline asynchronous learning, real-time online activitiesProfessional background of instructorsSelf-directed learningDiscussionCollective learning, individual learning, group learningFormative and summativeSelf-assessmentObjective evaluationLearner and instructor feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsChang L et alThe United States20262024N/AN/AN/ASchool educationN/AKnowledge, skillClinical application, ethical, legal, communication, collaboration, healthcare quality improvementCourseIndependent, IntegratedOptionalModular, spiralPreclinical theoretical, clinical clerkshipN/AOnline asynchronous learningN/AN/ADiscussion, case, hands-on practiceN/AN/ATeacher evaluationSubjective report, behavior evaluationN/AN/AN/ALindqwister et alThe United States2021N/A<60N/AClinical experts, AI specialists, or authoritative organizationsPostgraduate medical educationN/AKnowledgeN/ACourseIndependentCompulsoryModularN/AMore than one monthFace-to-faceProfessional background of instructorsDidactic teaching, self-directed learningDemonstrationCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, strengthening instructional interactionJalal AHB et alThe United Kingdom20242022<60N/AClinical expertsSchool education, continuing medical educationNeurosurgeryKnowledge, skillTechnical skillCourseIndependentOptionalModularN/ALess than one weekOnline asynchronous learning, real-time online activitiesN/AFlipped classroomDiscussion, demonstrationIndividual learning, group learningSummativeSelf-assessmentSubjective report, objective evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationPark SH et alThe United States20242022<60N/AClinical expertsSchool educationN/AKnowledge, skillCourseIntegratedOptionalModularPreclinical theoreticalN/AOnline and face-to-faceN/AFlipped classroomDiscussion, case, demonstrationCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AMihalas GI et alRomania2025N/A<60Questionnaire surveysClinical expertsSchool educationN/AKnowledgeClinical application, ethical, legalCourseIndependentN/AN/AN/AN/AN/AN/ADidactic teachingHands-on practiceCollective learningN/ATeacher evaluationObjective evaluationLearner and instructor feedback, other stakeholders’ feedbackCourse content, and curriculum organizationN/AAbid A et alThe United States20242021<60N/AN/ASchool educationN/AKnowledgeTechnical skill, clinical application, ethical, legalCourseIntegratedOptionalModularClinical clerkshipOne week to one monthOnline asynchronous learningProfessional background of instructorsSelf-directed learningProject-based learningIndividual earningSummativeSelf-assessmentSubjective report, behavior evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/APizzolla I et alBelgium20232022–2023 academic year<60N/ASchool educationN/AKnowledge, skillTechnical skill, clinical applicationCourseIndependentCompulsoryModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningDiscussion, case, hands-on practiceIndividual learningFormativeteacher EvaluationObjective evaluationLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, integrating AI into different learning phasesLindqwister et alThe United States2023N/A<60N/AClinical experts, AI specialists, or authoritative organizationsPostgraduate medical educationN/AKnowledgeN/AN/AIndependentOptionalModularN/AN/AOnline and face-to-faceProfessional background of instructorsDidactic teachingDiscussion, project-based learningCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedback, other stakeholders’ feedbackOverall course quality and learner satisfactionN/AHedderich DM et alGermany2021N/A90-150N/AN/ASchool education, postgraduate medical education, continuing medical educationN/AKnowledge, skillClinical application, ethical, legalCourseIndependentOptionalModularN/AMore than one monthOnline and face-to-faceN/AFlipped classroomDiscussion, demonstrationCollective learning, individual learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AAgarwal G et alThe United States2025N/A>150Literature reviewInterdisciplinary teamsSchool educationNursing, physical therapyKnowledgeClinical application, ethical, legalLectureIndependentN/AModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningCaseIndividual learningFormativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationStrengthening instructional interactionWang A et alNew Zealand2026N/AN/AN/Ainterdisciplinary teamsSchool educationRadiologyKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedCompulsoryModularPreclinical theoreticalN/AN/AProfessional background of instructorsDidactic teachingCaseCollective learningSummativeTeacher evaluationObjective evaluationLearner and instructor feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical components, implementing formal assessmentTaskiran N et alTurkey20232022>150N/AN/ASchool educationNursingKnowledgeClinical application, ethical, legalCourseIndependentOptionalN/APreclinical theoreticalMore than one monthFace-to-faceN/ADidactic teachingN/ACollective learningSummativeSelf-assessmentSubjective reportN/AOverall course quality and learner satisfactionN/ATejani AS et alThe United States20242021<60Questionnaire surveysClinical expertsPostgraduate medical educationRadiologyKnowledge, skillTechnical skill, clinical application, ethical, legalLectureIntegratedOptionalModularN/AMore than one monthFace-to-faceN/ADidactic teaching, flipped classroomDiscussion, case, hands-on practice, project-based learningCollective learning, group learningN/AN/AN/ALearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/AVan Kooten MJ et alThe Netherlands2024N/A<60N/AInterdisciplinary teamsPostgraduate medical educationRadiologyKnowledgeClinical application, ethical, legalCourseIndependentCompulsoryModularN/ALess than one weekN/AProfessional background of instructorsDidactic teachingDiscussion hands-on practiceCollective learning, group learningSummativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationEnhancing practical componentsLaupichler MC et alGermany20222021<60N/AClinical expertsSchool educationN/AKnowledgeClinical applicationCourseIndependentOptionalModularClinical clerkshipMore than one monthOnline asynchronous learning, real-time online activitiesProfessional background of instructorsFlipped classroomDiscussion, hands-on practiceCollective learning, individual learningFormative and summativeSelf-assessmentSubjective reportLearner feedbackOverall course quality and learner satisfaction, course content, and curriculum organizationN/ALevingston H et alThe United States2024N/AN/AN/AN/ASchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIndependentN/AModularPreclinical theoreticalLess than one weekOnline asynchronous learningN/ASelf-directed learningN/AIndividual learningFormativeSelf-assessmentObjective evaluationN/AOverall course quality and learner SatisfactionN/ASchaaf J et alGermany20252025N/AN/AInterdisciplinary teamsSchool education, continuing medical educationN/AKnowledgeClinical application, ethical, legalCourseIndependentN/AModularN/ALess than one weekOnline asynchronous learningN/ASelf-directed learningCase, storytellingIndividual learningFormative and summativeSelf-assessmentObjective evaluationN/AN/AN/AYangöz ŞT et alTurkey2026202490-150N/AN/ASchool educationNursingKnowledge, skillClinical application, ethical, legalCourseIndependentCompulsoryModularN/AMore than one monthFace-to-faceProfessional background of instructorsDidactic teachingDiscussion, hands-on practice, demonstrationN/AFormative and summativeTeacher evaluationObjective evaluationLearner feedbackN/AIntegrating AI into different learning phasesDonkor A et alGhana2025N/A90-150N/AInterdisciplinary teamsSchool educationMedical imagingKnowledgeTechnical skill, clinical application, ethical, legalLectureIndependentOptionalModularPreclinical theoretical, clinical Clerkshipless than one weekFace-to-faceN/ADidactic teachingDiscussionCollective learning, group learningFormativeSelf-assessmentSubjective reportN/AN/AN/AHu R et alCanada20222019>150N/AInterdisciplinary teamsSchool education, postgraduate medical educationN/AKnowledge, skillTechnical skill, clinical applicationWorkshopIndependentOptionalModularPreclinical theoreticalmore than one monthOnline and face-to-faceN/ADidactic teachingDiscussion, case, hands-on practiceCollective learningN/AN/AN/ALearner and instructor feedback, other stakeholders’ feedbackN/AEnhancing practical componentsAlZaabi A et alOman2025N/AN/AN/AN/ASchool educationN/AKnowledge, skillTechnical skill, clinical application, ethical, legalCourseIntegratedN/ASpiralPreclinical theoreticalN/AN/AN/AFlipped classroomDiscussion, project-based learningGroup learningSummativeSelf-assessmentSubjective reportN/AN/AN/AAbuadas M et alJordan20252024<60N/AInterdisciplinary teamsSchool educationNursingKnowledgeClinical application, ethical, legalCourseIndependentOptionalModularClinical clerkshipLess than one weekN/AN/AN/ADiscussion, case, project-based learningCollective learning, group learningSummativeSelf-assessmentSubjective reportN/AN/AN/A Analysis of Medical AI Education Programs Characteristics of Included Studies A total of 14 studies were published in the United States;24–37 4 in Germany;38–41 2 in Canada;42,43 2 in Turkey;44,45 2 in the United Kingdom;46,47 and 1 each in Australia,48 Belgium,49 Egypt,50 Ghana,51 Iran,52 Japan,53 New Zealand,54 Romania,55 Saudi Arabia,56 the Netherlands,57 Oman,58 and Jordan.59 By publication year, 2 articles were published in 2021,31,39 2 in 2022,40,43 6 in 2023,25,26,34,42,44,49 7 in 2024,28,32,33,36,37,47,57 13 in 2025,27,29,35,38,41,46,50–52,55,56,58,59 and 6 in 2026.24,30,45,48,53,54 Reference 24 described both a beginner-level AI course and an advanced data analysis course; only the beginner program was retained for analysis. References 29 and 36 covered both introductory and advanced curricula, which were integrated into a single coding unit during data extraction. The curriculum series in Reference 30 incorporated elective modules reported in Reference 26. Reference 34 represented a revised iteration of the curriculum originally presented in Reference 31. Although Reference 41 had not been implemented at the time of publication, its launch was scheduled for January 2025, and it was included in the final analysis after team discussion. Course Delivery Information 23 studies explicitly documented the year of first program delivery: 2 programs launched in 2019,26,43 3 in 2021,33,36,40 7 in 2022,25,27,32,42,44,47,50 2 in 2023,28,52 6 in 2024,29,30,45,48,56,59 and 2 in 2025,41,53 1 during the 2022–2023 academic year.49 Five programs confirmed that their programs had been offered in multiple rounds.26,33,43,48,50 Twenty‑eight programs specified the number of enrollments: 17 programs had fewer than 60 participants per iteration,25,26,28,31–34,36,38,40,42,47–49,55,57,59 5 had 90–150 participants,29,39,45,51,53 and 6 enrolled more than 150 learners.27,35,43,44,50,56 Four studies specified withdrawal rates,25,27,39,40 with the lowest being 9.38%40 and the highest 73.33%.25 It should be noted that the enrollment numbers for programs offered in multiple rounds were calculated based on the total number of students. Curriculum Development 26 of the included studies explicitly documented their curriculum development methodologies. Seven employed a narrative literature review approach, primarily to identify performance gaps, define learning objectives, select teaching content, and establish assessment methods.24,25,28,35,50,52,56 Six studies adopted questionnaire surveys, which were used to determine learning needs and curriculum scope through input from educators, students, domain experts, and professional organizations.24,25,27,36,52,55 Thirteen programs were developed by interdisciplinary teams consisting of clinical specialists and AI experts.26,29,35,38,41–43,46,51,52,54,57,59 Twelve programs were designed solely by clinical domain experts.27,28,31,32,34,36,40,47,48,50,55,56 Among these 12, 4 were developed by clinical experts with established AI literacy;28,32,47,48 3 had their content reviewed by external AI specialists or authoritative organizations;31,34,50 and 2 were developed under the supervision of professional bodies or interdisciplinary working groups.27,36 One program was independently designed and developed by AI experts alone.24 Target Audiences Twenty-seven studies focused on university-based medical education.25–28,30,32,33,35,37–41,43–47,49–55,58,59 Of these, 11 targeted undergraduate medical students,39,43–47,49–51,54,59 one involved master’s-level medical trainees,47 and 8 were designed for doctoral medical students.25,26,28,30,32,33,37,39 The remaining 9 could not be clearly categorized due to vague reporting or cross-national differences in educational systems. Nine programs belonged to postgraduate medical education,25,27,31,34,36,39,42,43,57 mainly serving resident physicians. Seven fell within continuing medical education,24,27,39,41,47,48,56 targeting practicing clinicians, nursing staff, and medical researchers. Sixteen studies specified clear disciplinary orientations: 6 in nursing,28,35,44,45,56,59 5 in radiology,36,42,46,54,57 and 2 in neurosurgery.25,47 Single studies were identified in pediatrics,48 medical imaging,51 and physical therapy.35 Six programs explicitly stated prerequisite prior knowledge for enrollment.25,26,29,31–33 Among these, References 26 and 32 required learners to complete pre-course self-study via assigned academic papers, instructional videos, or academic lectures. Learning Objectives In terms of curriculum learning objectives, the focus was primarily on the knowledge and skill domains. Fifteen studies focused on the knowledge domain,27,31,33–35,38,40,41,44,46,51,53,55,57,59 while another 21 covered both knowledge and skill domains.24–26,28–30,32,36,37,39,42,43,45,47–50,52,54,56,58 20 studies explicitly formulated their learning objectives in accordance with Bloom’s taxonomy.25,26,28–35,38,39,42,43,51–54,57,58 To assess the consistency of the two coders’ evaluations of the cognitive levels of the learning objectives, we employed the Kappa consistency test. The results yielded a Kappa coefficient of 0.93 (95% CI: 0.900–0.967), indicating an extremely high level of inter-rater agreement. The results showed that only 4 of these remained at the understanding level or below;32,51,54,57 the rest targeted higher-order cognitive skills, including applying, analyzing, evaluating, or creating. Specifically, 5 programs focused on the remembering level, 20 on understanding, 11 on applying, 12 on analyzing, 8 on evaluating, and 8 on creating (Table 2). To examine whether the learning objectives were uniformly distributed across the six classification levels, we employed a chi-square test. The results showed that χ2 = 164.24, p < 0.001, indicating a significant deviation from uniformity, with objectives predominantly concentrated at the understanding level.Table 2Distribution of Learning ObjectivesBloom’s Learning ObjectivesReferences (Number of Learning Objectives)Remembering28 (1), 38 (1), 42 (7), 52 (2), 58 (1)Understanding25 (5), 26 (3), 28 (1), 29 (1), 30 (4), 31 (2), 32 (16), 33 (6), 34 (3), 35 (4), 38 (2), 39 (2), 42 (9), 43 (1), 51 (25), 52 (2), 53 (2), 54 (5), 57 (3), 58 (5)Applying26 (2), 28 (1), 29 (3), 30 (1), 31 (2), 34 (1), 39 (2), 42 (6), 52 (1), 53 (1), 58 (2)Analyzing26 (5), 28 (1), 29 (2), 30 (5), 31 (2), 33 (1), 35 (2), 42 (1), 43 (1), 52 (1), 53 (1), 58 (2)Evaluating26 (4), 28 (1), 30 (6), 33 (1), 38 (2), 39 (1), 42 (3), 58 (4)Creating25 (1), 26 (11), 28 (1), 29 (2), 30 (3), 39 (1), 43 (1), 52 (1)Notes: Two coders independently extracted the verbs of the learning objectives and compared them with the verbs described in Bloom’s taxonomy. Matched verbs were retained and classified under the corresponding cognitive level, while discrepancies were determined by a third coder. “Remembering” corresponds to “28 (1)” in Reference 28, which represents a learning objective at the memory level of remembering. Table 2Distribution of Learning ObjectivesBloom’s Learning ObjectivesReferences (Number of Learning Objectives)Remembering28 (1), 38 (1), 42 (7), 52 (2), 58 (1)Understanding25 (5), 26 (3), 28 (1), 29 (1), 30 (4), 31 (2), 32 (16), 33 (6), 34 (3), 35 (4), 38 (2), 39 (2), 42 (9), 43 (1), 51 (25), 52 (2), 53 (2), 54 (5), 57 (3), 58 (5)Applying26 (2), 28 (1), 29 (3), 30 (1), 31 (2), 34 (1), 39 (2), 42 (6), 52 (1), 53 (1), 58 (2)Analyzing26 (5), 28 (1), 29 (2), 30 (5), 31 (2), 33 (1), 35 (2), 42 (1), 43 (1), 52 (1), 53 (1), 58 (2)Evaluating26 (4), 28 (1), 30 (6), 33 (1), 38 (2), 39 (1), 42 (3), 58 (4)Creating25 (1), 26 (11), 28 (1), 29 (2), 30 (3), 39 (1), 43 (1), 52 (1)Notes: Two coders independently extracted the verbs of the learning objectives and compared them with the verbs described in Bloom’s taxonomy. Matched verbs were retained and classified under the corresponding cognitive level, while discrepancies were determined by a third coder. “Remembering” corresponds to “28 (1)” in Reference 28, which represents a learning objective at the memory level of remembering. Distribution of Learning Objectives Notes: Two coders independently extracted the verbs of the learning objectives and compared them with the verbs described in Bloom’s taxonomy. Matched verbs were retained and classified under the corresponding cognitive level, while discrepancies were determined by a third coder. “Remembering” corresponds to “28 (1)” in Reference 28, which represents a learning objective at the memory level of remembering. Selection and Organization of Teaching Content Selection of Teaching Content All programs covered foundational AI knowledge, including core concepts, terminology, fundamental algorithms, alongside the technological evolution, practical functions, strengths, and limitations of AI. Eighteen programs incorporated technical skill training,24,25,27–29,33,36,37,42,43,47–49,51,52,54,56,58 focusing on operational proficiency and developmental workflows for AI tools. Thirty programs included applied clinical modules,24,26–30,33,35–46,48–52,54–59 covering AI applications in academic learning and routine healthcare settings. Twenty‑three programs addressed ethical and legal dimensions,24,25,28–30,33,35–37,39,41,44–46,51–59 outlining normative principles and potential ethical and legal challenges arising from AI adoption. Four programs covered communication,26,30,53,56 highlighting the changes of physician-patient communication methods with the participation of AI and communication approaches. One program integrated content on inter-professional collaboration and healthcare quality improvement30 (Table 3).Table 3Distribution of Course Content TopicsContent TopicFrequency, RateBasic Knowledge36, 100%The development history of AI; medical functions and roles of AI; current application status and trends; strengths and limitations24, 27, 33, 35, 38, 40, 41, 44, 45, 48-50, 52, 55, 56Basic concepts, terminology, functions, and working principles (algorithms)24‑37, 39, 40, 42‑55, 57‑59Basic Skill18, 50%Tool development skills (data processing, algorithm development, validation, and evaluation)24, 25, 27, 29, 33, 36, 42, 43, 47, 48, 51, 54AI tool usage skills (for example, prompt engineering)24, 28, 36, 37, 49, 52, 56, 58Application30, 83.33%Learning ApplicationsAI learning tools: strengths, limitations, and methods24, 30, 37Clinical Practice ApplicationsAI clinical application fields, strengths and weaknesses, and barriers26‑29, 35, 38‑41, 44, 46, 48‑50, 55, 56, 59AI Clinical Application Tools30, 40, 54, 56, 57Construction of AI clinical application solutions, deployment pathways, and their adaptability, feasibility, and performance evaluation26, 30, 33, 36, 38-40, 42, 43, 45, 46, 51, 52, 54, 57, 58Ethics and Law23, 63.89%Ethical and legal issues and governance in the application of AI24, 25, 28‑30, 33, 35‑37, 39, 41, 44‑46, 51‑59Cooperation1, 2.78%The transformation of collaborative practices among healthcare professionals and the development and application of collaborative strategies30Communication4, 11.11%The transformation of doctor–patient communication involving AI and the development and application of communication strategies26, 30, 53, 56Quality Improvement1, 2.78%Improvements in AI healthcare services and quality30Abbreviations: AI, artificial intelligence; PRISMA-ScR, the preferred reporting items for systematic reviews and meta-analyses extension for scoping reviews; ML, machine learning; DL, deep learning. Table 3Distribution of Course Content TopicsContent TopicFrequency, RateBasic Knowledge36, 100%The development history of AI; medical functions and roles of AI; current application status and trends; strengths and limitations24, 27, 33, 35, 38, 40, 41, 44, 45, 48-50, 52, 55, 56Basic concepts, terminology, functions, and working principles (algorithms)24‑37, 39, 40, 42‑55, 57‑59Basic Skill18, 50%Tool development skills (data processing, algorithm development, validation, and evaluation)24, 25, 27, 29, 33, 36, 42, 43, 47, 48, 51, 54AI tool usage skills (for example, prompt engineering)24, 28, 36, 37, 49, 52, 56, 58Application30, 83.33%Learning ApplicationsAI learning tools: strengths, limitations, and methods24, 30, 37Clinical Practice ApplicationsAI clinical application fields, strengths and weaknesses, and barriers26‑29, 35, 38‑41, 44, 46, 48‑50, 55, 56, 59AI Clinical Application Tools30, 40, 54, 56, 57Construction of AI clinical application solutions, deployment pathways, and their adaptability, feasibility, and performance evaluation26, 30, 33, 36, 38-40, 42, 43, 45, 46, 51, 52, 54, 57, 58Ethics and Law23, 63.89%Ethical and legal issues and governance in the application of AI24, 25, 28‑30, 33, 35‑37, 39, 41, 44‑46, 51‑59Cooperation1, 2.78%The transformation of collaborative practices among healthcare professionals and the development and application of collaborative strategies30Communication4, 11.11%The transformation of doctor–patient communication involving AI and the development and application of communication strategies26, 30, 53, 56Quality Improvement1, 2.78%Improvements in AI healthcare services and quality30Abbreviations: AI, artificial intelligence; PRISMA-ScR, the preferred reporting items for systematic reviews and meta-analyses extension for scoping reviews; ML, machine learning; DL, deep learning. Distribution of Course Content Topics Abbreviations: AI, artificial intelligence; PRISMA-ScR, the preferred reporting items for systematic reviews and meta-analyses extension for scoping reviews; ML, machine learning; DL, deep learning. Organization of Teaching Content In terms of delivery format, 25 took the form of formal courses,24–26,28,30–33,37,39–41,44,45,47,49,50,52–59 4 were delivered as lecture series,27,35,36,51 6 as workshops,29,38,42,43,48,53 and one as a seminar.46 Regarding integration with the established curriculum, 26 were designed as independent courses,24–27,29–31,34,35,37,39–41,43–45,47–52,55–57,59 while 11 integrated AI content into existing medical curricula.28,30,32,33,36,38,42,46,53,54,58 Reference 30 consisted of three components: an embedded integrated module, an unspecified middle section, and a final standalone module. In terms of course requirement status, 7 programs were compulsory,28,31,45,49,53,54,57 and 23 were offered as optional fered as optional electiveselectives.25–27,29,30,32–34,36,38–40,42–44,46–48,50–52,56,59 Teaching contents were organized modularly in 32 programs,24–43,45–51,53,54,56,57,59 and a spiral design in only 2.30,58 Among university-based programs, 19 clearly show the arrangement of the timetable, 11 were scheduled during the preclinical theoretical stage,30,32,35,37,43,44,49–51,54,58 and 10 during the clinical clerkship stage.25,26,30,33,38,40,46,51,53,59 Notably, References 30 and 51 spanned both theoretical and clinical learning phases. Course duration was reported in 26 studies: 13 lasted no more than 1 week,24,28,29,35,37,38,41,47,49,51,53,57,59 5 ran from 1 week to 1 month,25,26,33,42,50 and 8 extended beyond 1 month.31,36,39,40,43–45,56 Teaching Implementation Teaching Environment 13 programs adopted fully online asynchronous learning formats,24–27,30,33,35,37,40,41,47,49,50 among which 5 also offered real-time online discussions and other interactive teaching activities.26,27,40,47,50 Another 9 combined online and face-to-face instruction,29,32,34,38,39,42,43,52,56 with 5 delivering both in-person classroom sessions and live synchronous streaming for remote learners.29,32,34,42,52 The remaining 9 relied entirely on face-to-face teaching.28,31,36,44–46,48,51,53 12 programs specified the professional background of course instructors.28,29,31,33,34,40,42,45,48,50,54,57 With the exceptions of References 31 and 34, all other programs required instructors to be either AI specialists or teaching professionals with solid practical AI experience. Teaching Strategies Fifteen programs adopted traditional didactic teaching modes,28,31,34,36,43–46,48,51,53–57 with instructors delivering lectures either in person or via live online streaming. Thirteen programs employed self-directed learning approaches,24–27,29,31,33,35,37,41,49,50,53 enabling learners to study independently through watching instructional videos and reading extended learning materials. Nine programs implemented the flipped classroom model,32,36,38–40,42,47,52,58 in which learners completed self-paced online learning beforehand, followed by interactive discussions with faculty and peers, or teacher-guided inquiry-based learning activities. All programs incorporated lecture-based teaching. Twenty programs used seminar discussion sessions to facilitate interaction among learners, instructors, teaching assistants, and peers, to consolidate knowledge construction.27,30,32,34,36,38–40,43,45–47,49–52,56–59 15 programs adopted case-based learning to deepen learners’ understanding of AI-related knowledge and practical skills.24,26,28,30,32,35,36,38,41–43,46,49,54,59 12 programs arranged hands-on practical exercises to assist learners in applying theoretical knowledge and advancing their AI proficiency.24,25,28–30,36,40,43,45,49,55,57 10 programs applied demonstrative teaching to support learners’ knowledge acquisition.24,27,31,32,38,39,42,45–47 10 programs integrated project-based learning to foster inquiry-oriented study.25,26,33,34,36,38,52,53,58,59 One program each adopted role-playing,38 peer teaching,52 game-based learning,52 and storytelling.41 Twenty‑three programs organized teaching through whole-group collective learning.27–29,31,32,34,36,38–40,42–44,46,48,50,51,53–57,59 16 programs provided individual learning pathways.24–29,33,35,37–41,47,49,50 16 programs arranged collaborative small-group activities.25,26,31,32,34,36,38,47,50–53,56–59 Learning Outcome Assessment and Course Evaluation Learning Outcome Assessment Methods Six programs adopted formative assessment alone,24,26,35,37,49,51 20 programs employed summative assessment,27–29,31–34,38,39,42,44,46–48,52,54,56–59 and 6 programs combined both formative and summative assessment approaches.25,40,41,45,50,53 Twenty‑seven programs included self-assessment,24,27–29,31–35,37–42,44,46–48,50–53,56–59 while 7 programs involved teacher evaluation.25,26,30,45,49,54,55 Twenty‑one programs measured learning outcomes mainly through subjective reports including questionnaires and interviews.27–35,38–40,42,44,47,48,51,52,56–59 12 programs implemented objective evaluation via quizzes and formal examinations.24,26,37,41,45–47,49,50,53–55 Another 7 programs adopted behavior evaluation based on projects, assignments, reports, seminars, and reflective tasks.25,26,28,30,33,53,56 All programs except 624,36,37,41,45,55,58 clearly demonstrated positive impacts of AI education on learners’ AI knowledge, practical skills, and professional attitudes. Findings indicated that such programs strengthened learners’ understanding of core AI concepts, enhanced their capabilities in AI development and application, and improved confidence in reading AI-related literature and deploying AI tools in practice. Course Evaluation and Improvement Methods 24 programs specified course evaluation methods. Among these, 20 programs evaluated their curricula by collecting learner feedback through questionnaires, open-ended questions, and other survey methods.24–27,29,31–36,38–40,45,47,49,52,53,57 Four programs assessed course quality based on feedback from both learners and instructors.43,50,54,55 Three programs conducted iterative optimization and re-implementation based on feedback from multiple stakeholders—including learners and instructors—after initial course delivery.34,43,55 Nineteen programs reported overall high course quality and learner satisfaction.25–27,31–40,44,47,49,50,54,57 Of these, 18 programs evaluated the rationality of course content (such as coverage scope and learning materials) and curriculum organization (including course design and teaching activities).25,27,29,31–33,35,36,38–40,47,49,50,52,54,55,57 9 programs explicitly proposed increasing or enhancing practical components.27,31,43,46,49,50,52,54,57 Five programs suggested strengthening instructional interaction, such as increasing instructor guidance and small-group activities.25,29,31,35,46 Three programs recommended integrating AI content into different learning phases.26,45,49 Two programs advised implementing formal learning outcome assessment.24,54 Notably, learner assessments of knowledge and skills can reflect course effectiveness; however, this study has already analyzed these as learning outcome assessment methods, so no redundant analysis is provided here. Discussion Key Findings Structured programs and formal training are a critical approach to improving AI literacy of healthcare professionals. Currently, 17 countries, led by the United States, have developed medical AI curricula that are actively being taught, with significant growth in the last 5 years. More than 70% of the studies included in this review were published during the last 3 years. However, the overall extent of medical AI curriculum development and implementation remains limited. In terms of target audiences, most programs address medical students, while AI programs specifically designed for residents and practicing physicians are less common. Nursing and radiology have taken the forefront in developing medical AI courses, with other specialties making progress on this front. While many organizations and institutions have investigated medical AI education for medical professionals at different training levels and specialties, most programs remain in pilot, to our knowledge, with not many offered repeatedly. The enrollment number is relatively small, and some courses have high withdrawal rates, indicating that medical AI education is still in an early exploratory stage. In terms of curriculum development methods, the current medical education AI courses mainly rely on expert knowledge and experience for construction; the learning objectives mainly focus on knowledge and skills and involve high-level cognition, but the affective learning objectives are missing; the teaching content focuses on core content, clinical applications, ethics, and law, with limited focus on AI development or application skills, communication and collaboration, and quality improvement, which are mainly organized in a modular way and carried out in the form of elective courses, and strong content arrangement and integration are still required; various teaching methods are adopted to conduct online and in-person teaching activities for groups, teams, or individuals, but they lack solid theoretical foundations, and the instructional design team is insufficiently involved, and objective assessment of learning outcomes is also absent; course evaluation and improvement rely on subjective learner feedback, with limited use of instructor input and learning analytics. Program Development Methods This review systematically analyzed the approaches used to design educational programs. The results indicated that most curricula were developed by interdisciplinary teams that include AI specialists or by teachers who have some knowledge of AI. Designs are heavily influenced by individual expertise and experience, which is due to inconsistent competency standards for medical AI education, limited established curricula, and a lack of implementation. Most institutions and instructors set learning objectives independently, without formal assessments of future workforce skill needs or gaps between intended goals and learners’ current abilities. Literature reviews and surveys are the most common development strategies. These methods are widely adopted by other researchers. Tolentino et al12 performed a literature review to examine frameworks for medical AI education. Ki et al developed a South Korean curriculum based on survey data.60 The Association of American Medical Colleges surveyed AI courses offered in US and Canadian medical schools.61 Besides, the Delphi method is also an effective approach: Singla et al applied it to create a Canadian medical AI curriculum;62 Kern’s Six-Step Approach provides another structured path for the design and development of medical AI courses, such as Cai et al63 and Ahsan et al15, who used the framework to analyze and construct the curriculum framework. Curriculum teams may adopt these methods to design medical AI courses. Learning Objectives Learning objectives describe anticipated cognitive, psychomotor, and affective domains for learners. They serve as the primary basis for selecting course content, designing teaching strategies, and formulating evaluation methods for learning outcomes. Research results show that due to the different target audiences, the learning objectives of different courses or training also vary. However, all educational programs cover knowledge acquisition, with more than half of the courses involving skill teaching, and almost no program involves affective teaching, which indicates that the existing AI programs lack learning objectives for the emotional domain. In terms of the way of clarifying learning objectives, most programs frame objectives using Bloom’s Taxonomy, outlining the cognitive abilities for medical students needed to achieve the expected learning outcomes. Overall, the learning objectives are concentrated mainly at the understanding level, but most curricula also cover higher-order cognitive skills such as analyzing, evaluating, or creating. Medical learners need not only to understand the fundamentals of AI, but also to build, implement, and evaluate AI tools ethically. They must also adapt to evolving team dynamics and doctor-patient communication patterns shaped by AI integration.17,64–66 Miller’s Pyramid provides an alternative description of progressive medical competencies. Curriculum teams can use the model to plot expected proficiency levels, as Cai et al67 and others did to investigate healthcare AI competencies. Selection and Organization of Teaching Content Curriculum teams select and organize curriculum content that will help learners achieve learning outcomes. Existing programs always align curriculum contents with learning outcomes and assessment strategies, covering foundational AI knowledge, clinical applications, and ethical-legal principles. Fewer programs cover basic AI development and applied skills, and even fewer address collaboration, communication, or quality improvement. Most are standalone courses; few incorporate AI material into existing courses. They are offered mainly as electives during preclinical, clinical, or postgraduate programs for less than a month. Shorter formats reveal profound medical school workloads and busy clinical schedules—mandatory or long formats risk overburdening healthcare professionals.4 Most curricula describe a modular structure, while spiral progression remains rare. Alignment across learning stages and integration with existing medical content remains to be addressed. Program Implementation During delivery, course teams must consider instructional environments, resources, and teaching strategies. Current programs often combine multiple approaches. Self-directed learning promotes individualized study through online activities. Face-to-face sessions deliver traditional lectures, explain AI concepts, and demonstrate hands-on skills. Some courses use flipped learning: students complete independent study before engaging in guided discussions with instructors. Many programs also offer both in-person and live/recorded online sessions to accommodate diverse learning needs. Few programs explicitly base design on learning theories;26,30,50,53,54 most resorted to different methods of teaching didactic and practical design. Lectures remain the most common format of presentation. Case-based learning, discussions, hands-on activities, and project-based learning are other common themes. Such approaches promote interaction between learners, instructors, and peers, enhancing knowledge retention and skill development. Most programs require instructors to have AI experience, but instructional designers or learning science specialists are rarely involved Reference 28, 35. The use of dedicated design teams informed by learning theory has the potential to enhance the quality of online courses. Learning Outcome Assessment and Course Evaluation Learner self-report via survey dominates summative assessment. Although self-evaluations indicate improved knowledge, skills, and attitudes, limited formal and objective outcome measures have been demonstrated. Future work should develop reliable and valid assessment tools that will more accurately demonstrate learner competencies and measure objective progress. Existing medical AI curricula rely heavily on subjective student feedback to assess overall program quality, course content, and delivery implementation. Based on such input, common revisions include expanded faculty guidance and additional practical training components. Even so, insufficient attention is paid to instructor and institutional feedback, as well as empirical analysis of learners’ actual performance. The scalability and sustained implementation of medical AI education hinge directly on its instructional effectiveness. Curriculum developers should conduct ongoing evaluation and continuous course optimization while documenting tangible educational benefits. Such evidence allows institutional administrators to assess the merit and value of a course objectively and facilitates the formal incorporation of pilot programs into a medical training system. Strengths and Limitations This review adopted a systematic search strategy across major medical databases. It builds from broad literature coverage and methodologically reproducible procedures. Two independent researchers conducted study screening, data analysis, and qualitative coding to minimize potential research bias. There are several limitations to be acknowledged. The primary studies offered incomplete descriptions of course details, leaving important program descriptions unstated. As a result, qualitative coding and interpretive analysis were inevitably influenced by subjective judgment, which may result in minor discrepancies relative to the original training initiatives. The total pool of retrieved literature was small. Study methodological quality and publication quality were therefore not assessed during the screening process. Conclusion The continuous shift toward AI-enabled smart healthcare has brought forward new competency requirements for medical professionals. In recent years, AI-based clinical tools have advanced considerably, such as the nested ensemble deep learning model for predicting gynecological cancer risk,68 and the multimodal framework for the early diagnosis of Parkinson’s disease.69 These developments require medical staff not only to understand and apply AI systems, but also to interpret and evaluate the results they generate critically, and to communicate AI-informed decisions to patients. Targeted AI education is urgently needed to build AI proficiency of medical students, resident physicians, and practicing clinicians. This study systematically examined the design, development, and practice of existing medical AI education programs, providing references and inspirations for future curriculum construction and implementation of medical AI programs. All courses included in this study have been implemented in practice, rather than those still at the stage of theoretical construction. They thus provide a more accurate reflection of the current status and problems of existing programs and offer more operational guidance for the construction of medical AI courses. In addition, this study conducted a more in-depth analysis of these programs based on Bloom’s taxonomy. This systematic review indicates that medical AI education programs are still in the exploration stage, mainly consisting of small-scale pilot programs, and the design, implementation, and evaluation methods still need to be improved: affective learning objectives are missing; content related to AI development, application, communication, and collaboration remains insufficient; course scheduling, organization, and integration with existing courses need strengthening; teaching design teams are not adequately involved; and the learning results and course evaluation methods need to be perfected. To enhance the quality of the medical AI courses, we suggest that the relevant course teams give priority to the following areas. Firstly, AI applications, communication, and cooperation should be integrated as core teaching content into the course content. Medical students should not only understand the basic knowledge of AI, but also master the relevant skills for applying AI systems in clinical practice, thereby achieving intelligent healthcare. Secondly, the courses should introduce formal and verified learning outcomes and course evaluation tools, so as to more objectively measure the extent to which the medical students have achieved their learning goals and the quality of the courses, avoiding being influenced by subjective factors such as novelty effects. Finally, a longitudinal integrated AI course system should also be explored, integrating AI throughout the entire process of medical knowledge learning and clinical practice learning, so that medical students can progress from being beginners who understand AI knowledge to competent users who master the methods of using AI, and ultimately become experts leading the innovation of medical AI. We suggest that future research be conducted in the following three aspects. First, establish a standardized capability framework. Currently, the courses mainly rely on expert experience and lack a systematic framework for guidance. It is recommended to conduct surveys of interdisciplinary teams, including institutional administrators, medical experts, and AI experts, to construct a medical AI capability standard framework that reaches consensus among interdisciplinary experts. This will help institutions design and develop localized medical AI courses. Second, explore the effectiveness of different teaching models. The current teaching models are diverse but lack comparative effectiveness evidence. It is recommended to conduct comparative studies on learning outcomes of different teaching models and use subgroup interaction tests or moderator variable analysis to examine differences between different majors or training levels, and explore effective teaching strategies in different contexts. This will provide an effective theoretical basis for course teams to design teaching environments, select teaching strategies, and design teaching activities, thereby improving learning quality. Third, introduce data-driven assessment and improvement methods. Currently, the courses mainly rely on learners’ subjective feedback and lack objective indicators. It is recommended to introduce objective structured clinical examinations and other objective evaluation methods to measure learning outcomes and report effect sizes to enhance the comparability of the results; as the evidence base accumulates, it can be used to determine the best teaching strategies through network meta-analysis, determine the optimal teaching intensity through dose–response modeling, explore the impact of course characteristics on medical students’ completion through survival analysis, and monitor the learning process and measure learning outcomes through machine learning and other methods, promoting continuous improvement of evidence-based courses; employ decision analysis modeling methods such as Markov models and agent-based modeling methods to conduct cost–benefit analysis and barrier prediction for course implementation.