Sage Journals: Discover world-class research

Abstract

Generative artificial intelligence (GenAI) has introduced a transformative approach in medical informatics and education. AI-driven video models, such as Sora, HeyGen, Synthesia, and Google Veo 3, among others, can autonomously generate realistic clinical materials, including synthetic patients and simulated scenarios. This technology system represents an emerging domain of medical learning informatics that integrates AI-generated content, simulation, and pedagogy. This scoping review, based on selected studies, identifies and synthesizes educational outcomes, highlights the methodological limitations of AI-generated videos as training tools for medical students, and explores technical and pedagogical challenges to guide future research. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews framework, the literature searches were conducted across Google Scholar, PubMed, ScienceDirect, Scopus, and gray literature sources. Studies were included if they focused on AI-generated videos as educational tools for medical education. A single reviewer conducted screening of titles, abstracts, and full texts, and data were systematically extracted using a standardized charting form, including study design, AI tool utilized, outcomes, limitations, and challenges. Of the 970 retrieved records, 8 studies met the inclusion criteria. The latter demonstrated that AI-generated videos can enhance knowledge retention, skill acquisition, and learner engagement, outperforming traditional methods of delivering practical exercises in medical education. Reported challenges included issues with accuracy, limited emotional authenticity, ethical standards, and the necessity for pedagogical consistency. AI-driven videos are a promising innovation in medical education, offering scalable, interactive, and personalized learning. However, their integration requires a solid validation framework, interdisciplinary collaboration, and governance models that guarantee ethical and pedagogically appropriate use. Additionally, long-term, cross-institutional studies are necessary to evaluate the lasting educational and clinical effects.

Keywords

artificial intelligence generated-video medical informatics medical education virtual patients

Introduction

Background

Advances in generative artificial intelligence (GenAI) have reshaped the boundaries of digital learning and medical simulation. Among these innovations, artificial intelligence generated videos have emerged as a distinct technological and educational framework that warrants separate consideration. Also called AI-driven videos,¹ they are synthetic audiovisual learning materials automatically produced by GenAI models. They are typically generated through a process that uses a text or scientific term prompt (simple or multimodal).² AI-generated videos differ conceptually from virtual patients, which are interactive, software-based clinical scenarios designed for case-based training,³ and from traditional simulation modalities, such as real people, mannequins, or instructor-led simulation environments.⁴ They also differ from other AI-driven education tools, such as adaptive quizzes⁵ and intelligent tutoring systems that are similar to virtual patients.⁶ While these tools emphasize real-time feedback and decision-making, AI-driven videos offer pre-made, repeatable audiovisual scenarios. This enables students to observe, reflect, and practice skills independently, serving as a complementary method within interactive learning environments.

In medical informatics, AI-driven videos offer new possibilities for creating engaging, adaptive instructional materials that surpass traditional methods through immersive simulations. Video generation platforms such as Sora by OpenAI, HeyGen, Synthesia, and other multimodal systems can generate synthetic patient interactions, clinical cases, and procedural demonstrations. Creating videos with the aforementioned tools involves using prompts that simulate various clinical narratives,⁷ which are introduced in Sora or other AI video generators to produce realistic videos,⁸ featuring synthetic patients. This allows medical students to engage in interactive exercises to practice patient conversations in a digital environment.⁹ These text-to-video innovations¹⁰ integrate AI-based data synthesis with educational informatics to enable the automated creation, storage, and dissemination of training materials (Figure 1).

Figure 1.

Pathway for generating videos using an artificial intelligence driven solution.

From an informatics standpoint, AI-driven videos function as an intelligent media system that facilitates knowledge management, learner modeling, and simulation-based learning, shaping how students use tools¹¹ to improve their learning experiences.¹² These videos enable consistent replication of clinical encounters while minimizing reliance on human actors or expensive simulation setups. Through these scenarios, students interact with digital humans that convey realistic movements and expressions,^13,14 improving understanding, retention,¹⁵ self-confidence, and expertise.¹² This innovation is particularly relevant for medical education programs that are increasingly incorporating digital learning environments such as PCS,¹⁶ Geeky Medics,¹⁷ Healthy Simulation,¹⁸ Confluent,¹⁹ and adaptive instructional technologies.

Notably, these technologies are recognized for providing personalized,²⁰ self-directed,²¹ secure, and effective methods for students to enhance their skills,^22,23 while simultaneously delivering a realistic learning experience²⁴ that can strengthen confidence and foster self-efficacy.^25,26 Based on medical education principles, these AI-driven platforms can assume additional roles, such as (1) virtual simulation laboratories that effectively connect traditional classroom learning with authentic clinical experiences²⁶ and (2) realistic simulation scenarios for practitioners to rehearse intricate procedures and develop new skills without interrupting their clinical responsibilities.²⁷ Both roles can help reduce medical errors and improve patient outcomes.²⁸

However, the use of synthetic video content raises important issues. AI-generated videos may present downsides, such as inconsistent speech, distracting gestures, and emotional biases.¹⁵ In addition to the earlier issues, students report concerns regarding the depth of interpersonal connections, the authenticity of interactions, inaccuracies in content,^29,30 limited facial expressions, gestures, and monotone vocal expressions,³¹ the failure to compare with traditional methods, and varying levels of users’ technological skills.³²

To the author's knowledge, the number of primary studies on AI video generators for medical education remains limited; however, the affordability and availability of these technologies are increasing, owing to the emergence of numerous AI-driven platforms for medical training. Recent studies highlight several key reasons why AI isn’t widely used in medicine, despite its novelty, global accessibility, and low cost.³³ They highlight the existing gap between developers and end users,³⁴ the limited focus on AI-generated videos as teaching tools for medical students,³⁵ and their absence from educational curricula.³⁶ Together, these factors pose a challenge to the effective integration of AI-driven videos into the medical education system.

Objective

Based on the selected studies, this scoping review identifies and synthesizes educational outcomes, discusses methodological limitations associated with the use of AI-driven videos as a training tool for medical students, and explores technical and pedagogical challenges that could guide future research.

Methods

Study Design

This study was meticulously designed in accordance with the methodological framework proposed by Arksey and O’Malley,³⁷ and further refined by Leva et al.³⁸ The reporting of this study conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines.³⁹ The completed PRISMA-ScR checklist is provided as Supplemental File 1. A formal review protocol was not registered prior to conducting this study; however, predefined eligibility criteria, search strategies, and data charting procedures were established to promote transparency and reproducibility.

Protocol and Reporting Standard

The formal protocol was not registered prior to conducting the review. However, the review adhered to established scoping review methods, with eligibility criteria and analytic procedures set beforehand and applied consistently. Reporting complied with PRISMA-ScR guidelines to ensure transparency and reproducibility.

Eligibility Criteria

This research mainly investigated the use of AI-generated videos as educational tools for medical students. While it primarily focused on these students, it also included a study involving undergraduate nursing students because it examined AI-generated videos used in similar clinical education and simulation-based learning settings. Since health professions education shares pedagogical principles and overlapping skills, this inclusion aimed to provide a broader understanding of how AI-driven video technologies are used in clinical training. The findings regarding nursing students were carefully interpreted and not generalized beyond comparable educational contexts. Studies were included if they investigated the use of these tools in undergraduate, graduate, or continuing medical education; reported learning outcomes, limitations, and implementation challenges related to medical training; and were published in English in peer-reviewed journals or reputable gray literature sources.

The excluded articles were those that either did not discuss AI-generated video, mentioned it but did not target medical students, or referenced AI models without providing any medical context for students or incorporating AI-generated video. The characteristics of the studies, including the target participants, study objectives, AI tools used, and methodologies employed, were considered for inclusion. No restrictions were placed on publication year or geographical location to ensure comprehensive coverage of emerging research.

Information Source and Search Strategy

A systematic search was performed across 4 major databases: Google Scholar, PubMed, ScienceDirect, and Scopus. In addition to bibliographic databases, gray literature sources such as conference proceedings, preprint servers (eg, medRxiv and arXiv), and institutional reports were systematically searched to reduce publication bias. The identified records were screened using the same eligibility criteria as for peer-reviewed studies. Titles and abstracts were examined for relevance, followed by full-text review when sufficient methodological and educational details were available. Documents without clear authorship, transparent methodology, or educational relevance were excluded.

The study was conducted between February and December 2025, using a combination of vocabulary terms (such as MeSH) and free-text keywords related to AI, medical education, and AI-driven videos. The representative search string was “generative AI” OR “artificial intelligence” OR “text-to-video” OR “AI-driven video,” AND “medical education” OR “medical students” OR “synthetic patients” OR “simulation training.” Additionally, references within the selected studies were manually checked to identify further relevant sources.

Selection of Source Evidence

All records retrieved from databases were imported into EndNote (version 20, Clarivate Analytics) for reference management and duplicate removal. The author (PFI) screened the titles and abstracts of all records against predefined criteria. The full texts were screened thoroughly and simultaneously, with any uncertainties or conflicts addressed afterward. The screening process was documented and displayed in a PRISMA flow diagram (Figure 2).

Figure 2.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) flow diagram of study selection.

Due to resource and feasibility constraints, a single reviewer conducted the screening and study selection. This approach is acceptable in a scoping review, where the aim is to map available evidence rather than generate quantitatively pooled estimates. To reduce potential bias, predefined eligibility criteria were strictly applied, and uncertainties were reviewed multiple times during the screening process.

Data Charting and Extraction Process

Data extraction was conducted utilizing a standardized charting form specifically developed for this review. The data variables included: author(s) and publication year; study population and educational setting; type of AI-generated video tool used; study design and methodology; reported educational outcomes; and challenges and limitations as explicitly reported by the included studies. Data charting was also performed by a single reviewer based on feasibility, and the chart was checked for completeness and consistency before synthesis.

Data Synthesis and Analysis

All extracted data were organized narratively and thematically using an inductive approach. Studies were categorized based on their primary focus: (1) learning outcomes and (2) challenges and limitations issues. Descriptive analysis was used to determine the frequency of study characteristics, providing an overview of the distribution of evidence. Additionally, findings were examined to identify common themes in technological design, usability, interoperability, data ethics, and pedagogical integration. This dual perspective (educational and informatic) enables a comprehensive understanding of how AI-driven videos operate within digital learning environments in medical education.

Results

Selection of Sources of Evidence

The initial search identified 946 studies on AI-generated video in medical education, sourced from various databases, including 24 from gray literature. After removing duplicates (n = 150), 608 records were excluded based on title and abstract, leaving 212 records for eligibility assessment. Of these, 204 were excluded for specific reasons: 115 were not AI-generated video-based, 68 were not intended for medical students, and 21 were irrelevant because they referenced AI models without AI-generated video or any medical context. The remaining 8 records met the final inclusion criteria. The selection process is detailed in Figure 2, which follows the PRISMA flow diagram for scoping review.

Characteristics of Sources of Evidence

The included studies (n = 8) were published between 2022 and 2025 and researched the use of AI-driven videos in various medical education contexts. The vast majority of studies adopted quantitative methodologies^{31,40,41–45} employing either pre–post intervention or randomized control designs, while 1 study⁹ used a qualitative design focused on experiential feedback. Regarding objectives, the included studies examined learning effectiveness using pretest and posttest^40,42; assessed how tailored content captures focus⁴⁴; evaluated the realism of interactive communication training^9,41; and compared AI-generated interactive content with expert-led instruction,⁴³ with human-made video animation,⁴⁵ and explored perspectives on content creation and usability.³¹

The target populations included medical students,^{31,41,42,43,45} nursing students,⁴⁴ and trainees^9,40 at various training stages: from preclinical (first and second years) to postgraduate levels. Several studies have reported the use of AI-driven video tools, including Visla, HeyGen, Synthesia, CrazyTalk 8, Neuro VR, Google Veo 3, and multimodal systems such as Midjourney (Table 1). These tools facilitated the creation of synthetic patient simulations, interactive interviews, procedural demonstrations, and customized videos. Overall, the included studies demonstrate a growing diversity in how AI-generated video systems are designed and evaluated in medical training settings.

Table 1.

Characteristics of the Studies Included in the Scoping Review.

Author/Year	Population	Objectives	AI Tools Used	Methodology
(Verma, 2025)⁴⁰	Medical machine learning class	Effectiveness of AI-generated versus human-created educational videos on learning	Visla (AI-generative video platform)	Quantitative single group (pretest and posttest intervention)
(Althea et al, 2025)⁴⁴	Undergraduate nursing	Assess the limited focus on the specific educational needs	Not specific	Quantitative study
Kaya et al, 2025)⁴⁵	Second to fourth year medical students	To compare student and expert evaluations and preferences between AI-generated and human-made animated PBL triggers	Google Veo 3	Mixed-methods study (qualitative and quantitative)
(Chu, 2024)⁹	Medical trainees	Present a new educational tool designed to enable interactive, real-time simulations of challenging conversations through a video-based format using multimodal generative AI	Midjourney (version 5) Stable Diffusion XL LoRAs Pika Labs Runway Gen2 Stable Diffusion Video Lip-syncing-focused D-ID Synthetsia HeyGen	Qualitative (experience-based feedback) (control randomized comparative study)
(Lippitsch, 2024)⁴¹	Third and fourth Year medical students	Develop an interactive software system (ViPATalk) that allows students to train in posing questions to virtual patient avatars in free conversation.	ViPATalk CrazyTalk 8	Quantitative study (control randomized comparative study)
(Zhang, 2024)⁴²	Postgraduate medical students	This pilot study assesses the effectiveness of GenAI, specifically utilizing HeyGen, in producing personalized patient information videos in the Thai language.	HeyGen HeyGen's translation tool	Quantitative study Pretest and posttest (Likert scale survey-based)
(Fazlollahi, 2022)⁴³	First and second year medical students	To compare how VOA and remote expert instruction affect learners’ skill acquisition, as well as their affective and cognitive outcomes during surgical simulation training	NeuroVR (virtual reality) VOA	Quantitative study cross-sectional instructor-blinded randomized clinical trial
(Contreras, 2024)³¹	Medical students	To examine and assess AI video generation tools from a content creator's perspective.	Synthesia Colossyan Hour One HeyGen	Quantitative study survey-based

Abbreviations: AI, artificial intelligence; GenAI, generative AI; VOA, virtual operative assistant.

Results of Individual Sources of Evidence and Synthesis

Educational Outcomes

As shown in Table 2, the 8 studies included reported favorable educational outcomes from the use of AI-generated videos. These tools were found to be comparable to human-created videos in instructional effectiveness,^40,41,43 and were appreciated for their scalability, affordability, and learner engagement.^9,31,42 Specifically, AI-driven videos are more cost-effective than human-centered materials,⁴⁰ can simulate complex clinical scenarios,⁹ enhance comprehension and engagement,³⁴ and build medical skills in an unprecedented manner.⁴³ Overall, AI-driven video tools support traditional teaching methods, offering more opportunities for interactive, self-directed, and adaptive medical education.

Table 2.

Educational Outcomes Identified Across Studies.

Author/Year	Educational Outcomes
(Verma, 2025)⁴⁰	Artificial intelligence generated videos are just as effective as those created by humans. They offer a budget-friendly option and deliver content that maintains educational value.
(Althea et al, 2025)⁴⁴	AI-generated videos are a valuable supplementary instructional tool in nursing education.
(Kaya et al, 2025)⁴⁵	AI-generated triggers appear feasible, acceptable, and educationally promising.
(Chu, 2024)⁹	AI-generated videos have demonstrated their value in palliative care curricula by offering a scalable and high-quality simulation environment for practicing challenging conversations.
(Lippitsch, 2024)⁴¹	AI-generated videos were considered comparable to those produced through student role-play.
(Zhang, 2024)⁴²	AI-generated videos had a higher acceptance rate because they effectively boost understanding and improve educational methods.
(Fazlollahi, 2022)⁴³	AI-generated videos have proven effective in teaching surgical skills through simulation, creating an environment comparable to emotional and cognitive learning experiences.
(Contreras, 2024)³¹	AI-generated videos proved to be a practical alternative to traditionally made educational videos, offering an efficient and cost-effective way to produce educational content.

Identified Challenges and Limitations Across Studies

All studies, despite their findings, highlighted significant research gaps translated into technical and pedagogical challenges, as well as methodological limitations. Key themes of technical challenges included the necessity to (1) improve realism and emotional expressiveness in AI avatars⁹; (2) ensure AI-generated videos uphold high teaching and learning standards³¹; (3) ensure that the created tool align with the latest clinical guideline, evidence and best practices⁴⁰; (4) ensure accurate understanding and response to human speech and language utilizing AI⁴¹; (5) examine design factors such as video length, AI voice quality, and avatar fidelity⁴²; (6) explore hybrid methods that combine expert feedback with AI-generated content⁴³; (7) ensure that AI-driven videos are accurate and clinically reliable, aligning with international patient safety standards⁴⁴; and (8) develop evaluation criteria to compare AI-generated and human-created videos.⁴⁵ Furthermore, Table 3 outlines the methodological limitations and pedagogical challenges to inform future studies.

Table 3.

Pedagogical Challenges and Methodological Limitations to Inform Future Studies on Artificial Intelligence Driven Videos in Medical Education.

Author/Year	Methodological Limitations	Pedagogical Challenges
(Verma, 2025)⁴⁰	A decrease in the ability to identify meaningful differences in the results due to the limited generalizability of the findings across a diverse population.	Assess students’ critical thinking and problem-solving abilities in applying the knowledge acquired to new or complex cases.
(Alvarado et al, 2025)⁴⁴	Neglect of long-term student retention, skill transfer, and clinical application due to an immediate focus on short-term knowledge evaluation.	Ensure that the generated content maintains the same depth, nuance, and context as content created by humans.
(Kaya et al, 2025)⁴⁵	Immediate focus on short-term knowledge acquisition to foster skill retention, mastery, and practical application, which require long-term intervention.	Carefully evaluate if students are truly engaged with the content produced.
(Chu, 2024)⁹	Uncertainty in capturing the complexity, emotion, and unpredictability of genuine encounters, due to reliance on simulated, controlled conversational environments.	Recognize difficulties AI encounters in faithfully reproducing the complexity, nuance, and contextual sensitivity inherent in genuine clinical conversations.
(Lippitsch, 2024)⁴¹	Lowly accurate alignment of the generated content with questioning styles and spontaneous students’ responses due to dependence on speech recognition and natural language processing technologies.	Recognize that AI has difficulty understanding and responding appropriately to a wide range of students’ questioning styles (clinical reasoning and communication patterns).
(Zhang, 2024)⁴²	Uncertainty on determining whether the learning outcomes are influenced by the generated video or by the quality of the content itself.	Assess students’ willingness to engage with AI-generated content to avoid fewer benefits for those who are skeptical, regardless of the instructional quality of the tool.
(Fazlollahi, 2022)⁴³	Focusing on simulated surgical environments might not accurately represent actual operating conditions, such as stress levels, teamwork dynamics, and intraoperative variability.	Assess whether the generated content's feedback lacks depth, nuance, and contextual judgment to determine the need for individualized guidance for students.
(Contreras, 2024)³¹	Limited evidence of long-term impact due to the absence of longitudinal data demonstrating sustained learning outcomes and behavior change, which could lead to improvements in clinical practice.	Ensure that the generated content adheres to established pedagogical standards, including instructional coherence, clear objectives, and evidence-based information.

Moving forward, it will be vital to prioritize comprehensive validation, ethical considerations, and stakeholder collaboration to ensure the responsible implementation of this technology. Addressing these issues through careful instructional design, ethical review, and extensive long-term research is essential to guarantee that AI works as a valuable supplement rather than a replacement for human expertise in medical education.

Discussion

Summary of Evidence

This study identified 8 primary studies that address the use of AI-generated videos as learning tools in medical education. Collectively, these studies suggest that generative video technology may enhance learner engagement, conceptual understanding, and skill development, although current evidence remains preliminary.^15,46,47 Notably, the integration of such tools reflects a broader paradigm shift in medical informatics, in which AI is increasingly used to manage, personalize, and disseminate instructional content.

The reviewed evidence suggests that AI-generated video modules may complement traditional teaching by providing standardized and repeatable simulations, which are reported as potentially effective,^31,40,43 valuable,^9,44 comparable,⁴¹ and acceptable,^42,45 in early studies. This can help boost learners’ autonomy and enable self-paced study.^48–50 As McCloy foresaw, they may now significantly impact diagnostics and surgical solutions.⁵¹ Additionally, from an informatics standpoint, AI-driven videos are components of intelligent educational ecosystems that combine content-generation algorithms with user data analytics in training environments, aligning with key principles of digital learning in informatics, namely efficiency,⁵² accessibility,⁵³ and personalization.^54–56

Nevertheless, the finding underscores the persistent challenges in validating, governing, and ensuring ethical interoperability of such systems. Several studies have raised concerns about the realism of emotional expressions, the accuracy of medical information, and the transparency of underlying AI algorithms. These issues mirror broader concerns in medical informatics regarding algorithmic bias, data governance, and explainability. Moreover, the absence of a unified pedagogical framework for integrating AI-generated content hinders the systematic adoption of this approach in medical education.

To bridge these gaps, medical educators and informatics experts should collaborate to codesign AI video platforms that align with learning goals, accreditation requirements, and ethical principles. In particular, informatics should prioritize developing evaluation metrics for AI-driven videos, including performance analytics, cognitive load measures, and user engagement indicators. Incorporating AI-driven video systems into institutional learning management systems and electronic health record simulators could also improve authenticity and data interoperability.

Additionally, this involves improving the stakeholders’ literacy in AI tools,^41,43 simulating authentic clinical histories or lived experiences,^9,44 adhering to AI innovations that may improve quality and effectiveness,³¹ enhancing video features for better usability,⁴² and increasing its application to boost acceptance.^42,45 Moreover, reviews highlight the importance of clinical educators acquiring prior knowledge of AI innovations before integrating them into training programs.⁵⁷ This includes embedding AI into both preclinical and clinical curricula,^58,59 investigating novel areas, and addressing potential risks associated with AI usage in medical education.⁶⁰ These risks include inaccuracies, overreliance, integrity issues, authenticity concerns, bias, privacy breaches, security vulnerabilities,⁶¹ and scalability issues arising from the misuse of AI-driven tools.

Limitations

The paucity of primary studies on AI-generated videos in medical education limits the scope of this review. The included studies differed primarily in methodology, sample size, and educational context, limiting the generalizability of the findings. Rapid advances in AI technologies may render some tools or findings obsolete, necessitating ongoing updates to the literature. The search was limited to English-language sources, potentially excluding relevant studies published in other languages. Furthermore, the review does not consider long-term outcomes, such as the sustained impact of AI-generated videos on clinical performance and patient outcomes. Differences in technological literacy among medical students and educators add another layer of complexity that the studies did not thoroughly address.

A key methodological limitation is that screening and data extraction were conducted by a single reviewer due to resource constraints. This may increase the risk of selection bias, omission of relevant studies, and data extraction errors, potentially affecting the review's reliability and reproducibility. Additionally, including gray literature and preprints might lower the overall certainty of the evidence, as these sources often lack formal peer review and frequently present preliminary findings that may change. However, their inclusion was deemed important to capture emerging evidence in this rapidly developing field and to reduce publication bias. Lastly, although a protocol is not mandatory for scoping reviews, the absence of a registered protocol may decrease transparency and make reproducibility more difficult.

Conclusions

This review indicates that AI-generated videos show promise as a supportive tool in medical education, particularly for enhancing simulation-based and self-directed learning. Preliminary evidence suggests that these technologies may contribute to improved knowledge retention, skill development, and student engagement. However, the current evidence base remains limited, with a small number of heterogeneous studies employing diverse methodologies and outcome measures. As a result, findings should be interpreted cautiously. Important challenges remain, including verifying the authenticity of simulated experiences, aligning technological capabilities with educational needs, and addressing ethical issues such as data privacy, bias, and governance. Achieving the potential of an AI-generated video tool will require collaboration across disciplines among educators, technologists, and ethicists.

Future research should prioritize large-scale, methodologically rigorous, and longitudinal studies to validate educational effectiveness, develop a standardized evaluation framework, and clarify best practices for responsible integration into medical curricula.

Supplemental Material

sj-docx-1-mde-10.1177_23821205261437352 - Supplemental material for Artificial Intelligence Generated Videos as Supportive Tools in Medical Education: A Scoping Review

Supplemental material, sj-docx-1-mde-10.1177_23821205261437352 for Artificial Intelligence Generated Videos as Supportive Tools in Medical Education: A Scoping Review by Pinto Francisco Impito in Journal of Medical Education and Curricular Development

Supplemental Material

sj-docx-2-mde-10.1177_23821205261437352 - Supplemental material for Artificial Intelligence Generated Videos as Supportive Tools in Medical Education: A Scoping Review

Supplemental material, sj-docx-2-mde-10.1177_23821205261437352 for Artificial Intelligence Generated Videos as Supportive Tools in Medical Education: A Scoping Review by Pinto Francisco Impito in Journal of Medical Education and Curricular Development

Footnotes

Acknowledgements

The author extends thanks to Professor José Azevedo and Vasco Cumbe for their valuable guidance on his doctoral thesis, whose insights motivated him to conduct this research, representing a year of dedicated effort.

ORCID iD

Pinto Francisco Impito

Author Contributions

The author conceived the study, designed the methodology, conducted the literature search and screening, performed data charting and synthesis, and drafted and revised the manuscript.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research received institutional support through b-on, funded by Fundação para a Ciência e a Tecnologia (FCT), Portugal, via Porto University, Faculty of Engineering. The funding had no involvement in the study design, data collection, analysis, interpretation, or manuscript preparation.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental Material

Supplemental material for this article is available online.

References

Vora

Kadam

Mohite

, et al. AI-driven video summarization for optimizing content retrieval and management through deep learning techniques. Sci Rep. 2025;15(1):4058. doi:10.1038/s41598-025-87824-9

Borges

Ladele

Cunha

, et al. Multimodal prompt engineering for multimedia applications using the GPT model. 2024.

von Zadow

Buron

Harms

Behringer

Sostmann

Dachselt

. SimMed: combining simulation and interactive tabletops for medical education. In: Proceedings of the SIGCHI conference on human factors in computing systems, Paris, France, 2013:1469–1478.

Sun

Jiang

Dong

Feng

Shuai

. The evolution of simulation-based medical education research: from traditional to virtual simulations. Heliyon. 2024;10(15):e35627. 10.1016/j.heliyon.2024.e35627

Ningsih

Atmazaki

Ramadhan

. AI-powered adaptive quizzing: enhancing personalized learning, student engagement, and performance in digital classrooms. Salud Ciencia Tecnol. 2026;6:2647. doi:10.56294/saludcyt20262647

Létourneau

Deslandes Martineau

Charland

Karran

Boasen

Léger

. A systematic review of AI-driven intelligent tutoring systems (ITS) in K-12 education. npj Sci Learn. 2025;10(1):29. doi:10.1038/s41539-025-00320-7

Mahmud

Dwivedi

Chow

BJW

. Exploring the integration of artificial intelligence in cardiovascular medical education: unveiling opportunities and advancements. Can J Cardiol. 2024;40(10):1946–1949. doi:10.1016/j.cjca.2024.06.014

Leiker

Gyllen

Eldesouky

Cukurova

. Generative AI for learning: investigating the potential of synthetic learning videos. arXiv preprint arXiv:230403784. 2023.

Chu

Goodell

. Synthetic patients: Simulating difficult conversations with multimodal generative AI for medical education. arXiv preprint arXiv:240519941. 2024.

10.

Artsi

Sorin

Glicksberg

Korfiatis

Nadkarni

Klang

. Application of AI generated text-to-video in medical education: systematic review. medRxiv. 2025:2025.02.03.25321572. doi:10.1101/2025.02.03.25321572

11.

Benesch

Moore

Breyre

, et al. Primary palliative care education in emergency medicine residency: a mixed-methods analysis of a yearlong, multimodal intervention. Glob J Emerg Care. 2022;6(6):e10823. 10.1002/aet2.10823

12.

Pellas

. The impact of AI-generated instructional videos on problem-based learning in science teacher education. Educ Sci. 2025;15(1):102.

13.

Wang

Pan

Yan

Luan

. A survey on ChatGPT: AI-generated contents, challenges, and solutions. IEEE Open J Comput Soc. 2023;4:280–302.

14.

Simon

Chu

AJG

. Synthetic patients: simulating difficult conversations with multimodal generative AI for medical education. 2024.

15.

Netland

von Dzengelevski

Tesch

Kwasnitschka

. Comparing human-made and AI-generated teaching videos: an experimental study on learning effects. Comput Educ. 2025;224:105164. 10.1016/j.compedu.2024.105164

16.

[Internet]. Rhs. PCS. Accessed July 30, 2025. https://www.pcs.ai

17.

[Internet]. VPS. Geeky Medics. Accessed June 11, 2025. https://geekymedics.com/virtual-patient-simulator/

18.

Carn-Bennett

Artificial intelligence in healthcare simulation [Internet]. Healthy Simulation. Accessed August 17, 2025. https://www.healthysimulation.com/artificial-intelligence-healthcare-simulation/

19.

[Internet]. CETwAPA. Confluent. Accessed July 28, 2025. https://www.confluent.io/use-case/virtual-medical-training-platform/

20.

Thesen

Park

. A generative AI teaching assistant for personalized learning in medical education. npj Digit Med. 2025;8(1):627. 10.1038/s41746-025-02022-1

21.

Patil

Kou

Baptista-Hon

Monteiro

. Artificial intelligence in medical education: a practical guide for educators. MedComm – Future Med. 2025;4(2):e70018. 10.1002/mef2.70018

22.

Scherr

Halaseh

Spina

Andalib

Rivera

. ChatGPT interactive medical simulations for early clinical education: case study. JMIR Med Educ. 2023;9:e49877. doi:10.2196/49877

23.

Liu

Jin

, et al. From recorded to AI-generated instructional videos: a comparison of learning performance and experience. Br J Educ Technol. 2024;56(4):1463–1487. 10.1111/bjet.13530

24.

. Application of artificial intelligence generated content in medical examinations. Adv Med Educ Pract. 2025;16:331–339. 10.2147/AMEP.S492895

25.

Clarke

. Virtual reality simulation-the future of orthopaedic training? A systematic review and narrative analysis. Adv Simul (Lond). 2021;6(1):2. doi:10.1186/s41077-020-00153-x

26.

Desai

Mirza

. Virtual simulation skill labs and AI in medical education and healthcare. Med J Dr DY Patil Vidyapeeth. 2024;17(4):495–498.

27.

Lateef

. Simulation-based learning: just like the real thing. J Emerg Trauma Shock. 2010;3(4):348–352. doi:10.4103/0974-2700.70743

28.

Al-Elq

. Simulation-based medical teaching and learning. J Fam Commun Med. 2010;17(1):35–40. doi:10.4103/1319-1683.68787

29.

Kang

. The use of virtual patients for breaking bad news: a rapid review. Clin Teach. 2024/04/01 2024;21(2):e13681. 10.1111/tct.13681

30.

Ning

Ong

JCL

Cheng

, et al.

How can artificial intelligence transform the training of medical students and physicians?

Lancet Digit Health. 2025;7(10):100900. doi:10.1016/j.landig.2025.100900

31.

Contreras

Hossfeld

de Boer

Wiedler

Ghidinelli

. Revolutionising faculty development and continuing medical education through AI-generated videos. Journal of CME. 2024;13(1):2434322. doi:10.1080/28338073.2024.2434322

32.

Artsi

Sorin

Glicksberg

, et al. Application of AI generated text-to-video in medical education: systematic review. medRxiv. 2025:2025.02. 03.25321572.

33.

Cook

. Creating virtual patients using large language models: scalable, global, and low cost. Med Teach. 2025;47(1):40–42. doi:10.1080/0142159X.2024.2376879

34.

Zhang

Cai

Lee

Evans

Zhu

Ming

. AI in medical education: global situation, effects and challenges. Educ Inf Technol. 2024;29(4):4611–4633.

35.

Varma

Fernando

Ting

Aamir

Sivaprakasam

. The global use of artificial intelligence in the undergraduate medical curriculum: a systematic review. Cureus. 2023;15(5):e39701. doi:10.7759/cureus.39701

36.

Busch

Hoffmann

Truhn

, et al. Global cross-sectional student survey on AI in medical, dental, and veterinary education and practice at 192 faculties. BMC Med Educ. 2024;24(1):1066. doi:10.1186/s12909-024-06035-4

37.

Arksey

O'Malley

. Scoping studies: towards a methodological framework. Int J Soc Res Methodol. 2005;8(1):19–32. doi:10.1080/1364557032000119616

38.

Levac

Colquhoun

O'Brien

. Scoping studies: advancing the methodology. Implement Sci. 2010;5(1):69. doi:10.1186/1748-5908-5-69

39.

Tricco

Lillie

Zarin

, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med. 2018;169(7):467–473. doi:10.7326/M18-0850

40.

Verma

Leddo

. Comparing the effectiveness between human-generated videos and AI-generated videos on learning. 2025.

41.

Lippitsch

Steglich

Ludwig

, et al.

Development and evaluation of a software system for medical students to teach and practice anamnestic interviews with virtual patient avatars.

Comput Methods Programs Biomed. 2024;244:107964. doi:10.1016/j.cmpb.2023.107964

42.

Zhang

Lucas

Bem-haja

Pedro

. The effect of student acceptance on learning outcomes: AI-generated short videos versus paper materials. Comput Educ: Artif Intell. 2024;7:100286.

43.

Fazlollahi

Bakhaidar

Alsayegh

, et al. Effect of artificial intelligence tutoring vs expert instruction on learning simulated surgical skills among medical students: a randomized clinical trial. JAMA Netw Open. 2022;5(2):e2149008. doi:10.1001/jamanetworkopen.2021.49008

44.

Althea Monica

PAR

Joshua Antonio

Julius Francis

, et al. Evaluating the effectiveness of AI-generated health educational videos on nursing students’ knowledge acquisition of the International Patient Safety Goals (IPSG). J Interdiscipl Perspect. 2025;3(9). 10.69569/jip.2025.565

45.

Abdullah Bedir Kaya

YSK

Coşkun

Budakoğlu5

Iİ

. Comparison of AI-generated and human-made animated videos for medical education: experts and students preferred AI over humans. Rev Españ Educ Méd. 2025;4:677591. 10.6018/edumed.677591

46.

Bonanno

Kiraly

Siegel

Brasel

Cook

. Surgical palliative care training in general surgery residency: an educational needs assessment. Am J Surg. 2019;217(5):928–931. 10.1016/j.amjsurg.2019.01.008

47.

Whittaker

Letheren

Mulcahy

. The rise of deepfakes: a conceptual framework and research agenda for marketing. Australasian Market J. 2021;29(3):204–214. doi:10.1177/1839334921999479

48.

Schmelzer

. How AI is shaping the future of education [Internet]. Forbes. Accessed July 22, 2025. https://www.forbes.com/sites/ronschmelzer/2024/05/28/how-ai-is-shaping-the-future-of-education/

49.

Yildirim

Camci

Aygar

. Advancing self-directed learning through artificial intelligence. 2023:146–157. doi:10.4018/978-1-6684-6772-5.ch009

50.

Wang

Bonk

. Understanding self-directed learning in AI-assisted writing: a mixed methods study of postsecondary learners. Comput Educ Artif Intell. 2024;6:100247. 10.1016/j.caeai.2024.100247

51.

McCloy

Stone

. Science, medicine, and the future. Virtual reality in surgery. Br Med J. 2001;323(7318):912–915. doi:10.1136/bmj.323.7318.912

52.

Çeken

Taşkın

. Multimedia learning principles in different learning environments: a systematic review. Smart Learn Environ. 2022;9(1):19. doi:10.1186/s40561-022-00200-2

53.

Rocha Lourenço

Almeida

Oliveira

Paiva

. The accessibility of Digital Learning Resources: an analysis of audiovisual resources. 2022.

54.

Cao

Cui

, et al. A comprehensive exploration of personalized learning in smart education: from student modeling to personalized recommendations. ArXiv. 2024;abs/2402.01666.

55.

Yaseen

Mohammad

Ashal

Abusaimeh

Ali

Sharabati

A-AA

. The impact of adaptive learning technologies, personalized feedback, and interactive AI tools on student engagement: the moderating role of digital literacy. Sustainability. 2025;17(3):1133.

56.

Hoppe

Majumdar

Ogata

. Personalized learning environments—core concepts, technologies and practices. Inform Technol Educ Learn. 2024;4:Inv-p002. doi:10.12937/itel.4.1.Inv.p002

57.

Bowers

Graydon

Ryan

Lau

Tomlin

. Artificial intelligence-driven virtual patients for communication skill development in healthcare students: a scoping review. Aust J Educ Technol. 2024;40(3):39–57. 10.14742/ajet.9307

58.

Hale

Alexander

Wright

Gilliland

. Generative AI in undergraduate medical education: a rapid review. J Med Educ Curr Dev. 2024;11:23821205241266697. doi:10.1177/23821205241266697

59.

Russell

Lovett Novak

Patel

, et al. Competencies for the use of artificial intelligence-based tools by health care professionals. Acad Med. 2023;98(3):348–356. doi:10.1097/acm.0000000000004963

60.

Gordon

Michelle

Ajiboye Aderonke

, et al. A scoping review of artificial intelligence in medical education: BEME Guide No. 84. Med Teach. 2024;46(4):446–470. doi:10.1080/0142159X.2024.2314198

61.

Janumpally

Nanua

Ngo

Youens

. Generative artificial intelligence in graduate medical education. Sec Healthcare Professions Educ. 2025;11:1525604. 10.3389/fmed.2024.1525604

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.09 MB