Keywords
medical education, information technology, artificial intelligence, simulation, personalized learning, curriculum reform, digital transformation
-
Views
0
Downloads
How to cite this article
Shaban S and Magzoub ME. Artificial Intelligence in Medical Education and Assessment: The next step in the IT Revolution [version 1; peer review: awaiting peer review]. F1000Research 2025, 14:1360 (https://doi.org/10.12688/f1000research.173611.1)
NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.
Export Citation
Sciwheel
EndNote
Ref. Manager
Bibtex
ProCite
Sente
▬
✚
Review
Artificial Intelligence in Medical Education and Assessment: The next step in the IT Revolution
[version 1; peer review: awaiting peer review]
PUBLISHED 05 Dec 2025
OPEN PEER REVIEW
REVIEWER STATUS
This article is included in the AI in Medicine and Healthcare collection.
Abstract
Corresponding authors:
Sami Shaban, Mohi Eldin Magzoub
Competing interests:
No competing interests were disclosed.
Grant information:
The author(s) declared that no grants were involved in supporting this work.
Copyright:
© 2025 Shaban S and Magzoub ME.
This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
How to cite: Shaban S and Magzoub ME. Artificial Intelligence in Medical Education and Assessment: The next step in the IT Revolution [version 1; peer review: awaiting peer review]. F1000Research 2025, 14:1360 (https://doi.org/10.12688/f1000research.173611.1)
First published: 05 Dec 2025, 14:1360 (https://doi.org/10.12688/f1000research.173611.1)
Latest published: 05 Dec 2025, 14:1360 (https://doi.org/10.12688/f1000research.173611.1)
1. Introduction
Medical education has invariably been shaped over the years by changing pedagogic philosophies, scientific discoveries and societal needs. Traditionally, its basis was the model of apprenticeship to the profession, where the learner was taught mainly by observation and practice under senior physicians. This has been formalized and standardized in the early part of the 20th century by the Flexner Report, which has stressed curricula-structure, strong scientific foundations, and hospital-based clerkship training (Norman, 2012). While the Flexner model is considered one of the milestones of professionalization of medicine, its focus on lectures, memorization, and patient contact has been increasingly deemed inadequate in preparing physicians to address the complexities of healthcare today.
In the last 30 years, the unprecedented flowering of Information Technology (IT) has introduced new modalities of teaching-learning which work alongside the traditional paradigm and often contradict it in some respects. As a first step, the effort was devoted to accessibility and the creation of e-learning platforms, learning management systems, and digital libraries that storing medical knowledge and making it accessible on a wide scale at lower costs (Cook et al., 2008). These tools were then made into interactive multimedia resources able to present complex information in interesting ways (Kononowicz et al., 2019). The adoption of IT in the educational sphere was accelerated following the COVID-19 pandemic enabling educational institutions to increase the adoption of online and blended learning methodologies (Papapanou et al., 2022).
Artificial Intelligence (AI), together with IT, is now emerging as a transformational agent in the medical education sector. Compared to previous digital technologies that largely situated themselves around the delivery of content, the potential offered by AI introduces adaptive, data-driven and interactive features which allow the delivery of highly personalized learning experiences (Rincón et al., 2025). Machine learning models can fill gaps in the knowledge, personalize learning, and even simulate patient interactions in a virtual environment. Intelligent tutoring systems have the potential to personalize the questions, explanations, and paths of learning for individual learners, maximizing efficiently and retention (Fazlollahi et al., 2022). This ability is particularly important in medical education, where knowledge, procedural skills, and judgment are all of equal importance, and hence must be learned concurrently.
Medical schools and accreditation organizations have begun to recognize the need to incorporate digital competencies and AI literacy into medical school curricula (Car et al., 2025). An article from Harvard Medical School highlights that curriculum changes are becoming more focused on how medicine can be taught and also on how future physicians can become experts in how they can work with AI systems in clinical practice (Succi et al., 2025).
Despite the optimism for the future, there are challenges in the inclusion of IT and AI in medical education. These are barriers around cost and infrastructure, especially in resource-limited settings; faculty readiness must be addressed as educators need to be oriented to use these tools effectively and safely; ethical considerations like privacy of data and misuse by learners; concerns about AI algorithm bias and verification of results; and finally reducing the practice of medicine to a mechanized process (Chan & Zary, 2019; Zhui et al., 2024). There is also apprehensions among educators about faculty and learner over-reliance on AI that may undermine essential human qualities of medicine, such as empathy, intuition and interpersonal communication (Okamoto et al., 2025; Topol, 2023).
Nevertheless, there seems to be a consensus in the existing literature that the digital transformation of medical education is not an option but rather an imperative. With healthcare systems becoming more and more reliant on digital tools, from electronic health records to AI-powered diagnostic systems, medical education is having to change to equip medical students for a practice environment mediated by digital tools (Car et al., 2025; Khafizova et al., 2023; Prosen & Ličen, 2025).
This article performs a review focusing on mapping out the current literature on IT and AI in medical education and assessment, to synthesize current knowledge, identifying facilitators and obstacles, and drawing out implications of upcoming trends for the development of future curriculum and research. By doing so, it helps bring out the responsible use of these technologies to educate a new generation of physicians who are not only clinically competent but also technologically literate, ethically conscious, and able to adapt to an ever-changing healthcare landscape.
2. Methodology
We undertook a focused, narrative review designed for transparency rather than a systematic review. The aim of the review is to disclose the evolution of IT/AI and its utilization in different areas in medical education including instructional methods, curriculum design, student assessment, along with identification of benefits, challenges and ethical implications.
Database: Literature searches were conducted in PubMed, Google Scholar, and Scopus for studies published between January 2010 and December 2025 (last 15 years). The search strategy combined controlled vocabulary and free-text terms for three concept blocks: (Norman, 2012) medical/health-professions education (e.g., “medical education,” “health professions education”); (Cook et al., 2008) artificial intelligence and related technologies (e.g., “artificial intelligence,” “machine learning,” “deep learning,” “large language model*,” “intelligent tutor*,” “virtual patient*,” simulation, “virtual reality,” “augmented reality”); and (Kononowicz et al., 2019) educational foci (e.g., assessment, curriculum, OSCE, competency feedback). Titles/abstracts and then full texts were screened against predefined criteria. We included reviews (systematic, scoping, narrative), empirical studies (e.g., randomized and quasi-experimental trials, cohort and cross-sectional studies, qualitative work), and perspective pieces that directly addressed the use of information technology or artificial intelligence in medical education or assessment. We excluded reports of clinical AI without an educational application, opinion pieces with no educational focus, duplicates, and records without accessible full text unless central to the topic. After deduplication, evidence was synthesized narratively and thematically across five domains: e-learning/blended learning, VR/AR and simulation, AI-enabled tutoring/personalization, curriculum analytics, and AI-enabled assessment. Owing to heterogeneity in designs, interventions, and outcomes, no meta-analysis was planned. We considered study quality qualitatively by noting common threats to validity (e.g., selection bias, confounding, and measurement issues) and interpreted findings, accordingly, aligning the synthesis with the thematic structure outlined in the source narrative (see summary tables).
3. Results
3.1 The evolution of IT in medical education
The utilization of Information Technology (IT) in medical education has evolved through several stages from digitalization of static resources to the development of highly interactive and immersive learning environments (Delungahawatta et al., 2022). This evolution is part of bigger changes in pedagogy, healthcare needs, and technological prowess (Prober & Heath, 2012). Buja argues that medical education continues to change in response to scientific advances and societal needs (Buja, 2019). Following this progression, we can appreciate how digitization has gradually transformed the way in which healthcare workers are trained.
3.2 E-learning and blended learning
IT in medicine came first as e-learning and the creation of online content collections. They circumvented longstanding barriers to access through providing learners with electronic versions of textbooks, access to online databases, and video class demonstrations (Delungahawatta et al., 2022). Early systems were largely static, although they democratized the act of medical knowledge as quality resources became accessible to anyone, anywhere in the world, and often at a lower cost than using a print publication.
The idea of blended learning can be traced back to when institutions started to conceptually integrate digital resources with educational experiences of the traditional kind. Studies have demonstrated the efficacy of this model in terms of optimizing knowledge retention and learning capabilities through offering students the flexibility to learn theoretical material online at their own pace while using in-person time to facilitate applied, contextual, and clinical reasoning experience (Vallée et al., 2020). One of the prominent models of blended learning is the flipped classroom, which essentially flips the teaching-learning process by having learners access lecture information through online resources outside of the classroom and then spending the time in the classroom working on case-based discussion/problem-solving (Chen et al., 2017). Surveys conducted during and after the pandemic suggest that many institutions have decided that blended learning is more than a temporary fix and has become a permanent part of their curricular offering (Atwa et al., 2022; Seed et al., 2025; Wang et al., 2024a).
The benefits of e-learning and blended models do not end with their convenience. Evidence suggests that they: promote self-directed learning, which is a key competency for lifelong professional development; allow standardization of fundamental content, so educational quality is the same throughout institutions; and support scalability so that medical schools can find ways to meet an increased number of learners without burdening existing faculty proportionally (Lu et al., 2023).
3.3 Virtual reality and augmented reality
While the e-learning revolutionized the way content is delivered, Virtual Reality (VR) and Augmented Reality (AR) are the ones that have redefined experiential learning in the field of medicine. These technologies offer immersive environments where learners can practice with complex anatomical structures and simulations and practice procedures in a clinical setting, without any real-world risks.
Virtual Reality (VR) engulfs the learners in wholly digital environments (Kaggwa et al., 2025). VR has been widely applied in medical education in surgical training (Mao et al., 2021), visualization of anatomy (Zhao et al., 2020), and simulations of emergency response (Jung, 2022). For example, high-fidelity VR surgical simulators provide trainees the ability to repeatedly practice performing a procedure with real-time feedback on precision, efficiency, and handling of complications. Evidence has indicated that technical performance and confidence may be enhanced through VR-based surgical training before performance on real patients (Kyaw et al., 2019).
Augmented Reality (AR) is the application of digital information overlaying the physical world. This is a particularly effective application for anatomical education, as AR applications can project 3D models of organs or vascular systems directly onto cadavers or mannequins, for improved spatial understanding (Bogomolova et al., 2020). AR has also been applied to live clinical environments for projection of navigation guidance during surgical operations (Chien et al., 2022) or delivering visual middleware during investigations for patients (Wang et al., 2024b).
The pedagogical benefits of VR and AR relate to the fact that they provide a safe, repeatable, and very interactive opportunity to learn. Unlike traditional training, the emerging technologies allow learners to practice rare or high-risk scenarios (e.g., managing cardiac arrest, trauma) in a setting where patients are not at risk (Asoodar et al., 2024). They also aid procedural memory through the ability to memorize the technique of trial and error in controlled settings, a key factor in mastering surgical and diagnostic techniques.
Moreover, VR and AR also offer ways to share learning between people. Multi-user simulation allows learners at varying locations to communicate in the same virtual environment as they are working on clinical cases. This has been fairly handy in the case of global health education, for example, where institutions on different continents have been able to collaborate for the purpose of training (Lerner et al., 2020).
Despite all these advantages, the adoption of VR and AR is not without its significant problems. The technologies involved tend to be expensive, and there is a great deal of up-front capital to be invested in hardware, software and technical support. There is also a learning curve for both learners and educators, and getting simulations to truly reflect real-life clinical practice is also an ongoing challenge (Mergen et al., 2024). There are also some ethical considerations and the issue of distribution, how learners in resource-poor settings will be able to equally access and benefit from these innovations (Li et al., 2024; Raja & Al-Baghli, 2025).
3.4 From content to experience
Overall, the evolution of IT to AR/VR in medical education highlights a change in the conventional content delivery to total immersion and experiential learning. Earlier technologies were primarily about providing information, while the modern tools are conducive to interaction, simulation and actual engagement (Prober & Heath, 2012).
3.5 The advent of AI in medical education
Introduction of Artificial Intelligence (AI) in medical education is a great leap in the revolution of digital transformation of medical education teaching and learning. Unlike Information Technology (IT), which has been largely applied with the aim of enhancing access to, and delivery of learning materials, AI holds an edge in the sense that it has the capability of analyzing data, personalizing the learning and replicating human interaction with the learners (Chan & Zary, 2019). AI can be used to make medical education more personalized, effective, and consistent with the real-world clinical experience through natural language processing, machine learning, and prediction analytics (Gordon et al., 2024).
3.5.1 More student-centered learning and adaptive tutoring systems
One of the greatest efforts that AI can make is its capacity to offer personalized learning. Effective education-typical medical curricula (classified in the literature as conventional, routine or standardized) may not work with the singularity of and optimize the pace or deficit of individual learners. AI-powered platforms cannot only keep a check and balance on learner participation (Kassab et al., 2023), but they can also track the performance and areas that learners are having difficulties in real-time (Holderried et al., 2024). From this data, adaptive algorithms can recommend various types of readings, practice questions, or other explanations and ensure that learners are focusing on topics according to their needs (Monteverde-Suárez et al., 2024).
For example, an adaptive tutoring systems is being used in medical schools to help learners master the basic sciences through a variety of complex topics, such as physiology or pharmacology. These adaptive tutoring systems are like that of a human tutor in providing hints, explanations, and feedback to the learners, specifically to their level of knowledge (Abe et al., 2025). Evidence is available that adaptive tutoring systems are effective for both reaching retention and lessening cognitive overload, in contrast to conventional lecture-based learning. As a result, through constant and individualized assistance, AI can provide benefits not only in relation to the improvement of short-term performance but also in relation to the development of long-term professional competence (Fazlollahi et al., 2022).
3.5.2 AI in simulation and clinical training
AI has its role in moving from theoretical knowledge toward practical skills development. While high-fidelity simulations have been used in medical training for decades, the addition of AI ensures that the environment is not static and artificial (no pun intended), but dynamic and realistic, as well as reacting intelligently to learner actions (Cook et al., n.d.). Virtual patients powered by AI can mimic a broad range of physiological statuses & vital signs (Bray et al., 2019), speech (Holderried et al., 2024; Maicher et al., 2023), and even emotional instinct (Darnell et al., 2020).
Machine learning algorithms are used to evaluate learner performance and give feedback for diagnostic accuracy, treatment planning, and empathy in patient interaction (Schaye et al., 2022). This type of innovation is useful in promoting competency-based medical education (CBME), which involves a curriculum that concentrates on measurable skills and outcomes. By generating measurable performance data, simulations developed with an AI create a way for educators to better quantify competencies and for learners to mark their areas of weakness and strength themselves.
Another benefit is that advanced simulations can improve clinical preparedness. While there have long been virtual labs and case studies online that spur learning in IT, artificial intelligence-based simulations can understand and bring recommendations for dynamic, realistic patient encounters. These systems are used to give learners recurrent, risk-free training of infrequently seen medical presentations or complex physical assessments that may not be experienced on clinical rotations (Cook et al., n.d.). As a result, learners develop their diagnostic skills, the skills to make judgments and communication skills in safe but realistic environments. This leads to greater confidence and competency in making the transfer to the real world of practice.
Many of the challenges in the adoption of VR and AR remain with AI integration such as expensive technology, up-front capital, software and technical support, learning curve for both learners and educators, ethical consideration, and getting simulations to truly reflect real-life clinical practice (Mergen et al., 2024).
3.5.3 AI and data analytics for curriculum development
AI also has a significant role to play in the design and optimization of the curriculum. With medical knowledge more than doubling every few months, the challenge to educators is therefore to keep the curriculum up to date. AI can comprehend substantial literature about medicine, clinical guidelines, and research data to identify the gaps and new trends in training content (Buchanan et al., 2024; Harmsen et al., 2024; Waldvogel et al., 2025). For example, natural language processing (NLP) algorithms can be used to analyze medical databases highlighting underrepresented areas, making recommendations for curriculum change to account for shifting burdens of disease (Kavvadias et al., 2020).
Additionally, AI-supported learning analytics can help to identify patterns in cohorts to help understand the effectiveness of teaching methods or modules from the faculty’s standpoint. Such an evidence-based methodology can ensure the evidence-based, responsive, and aligned nature of the curricula based on the latest requirements of the healthcare field (Bojic et al., 2023; Komenda et al., 2015).
3.5.4 AI for improved assessment and feedback
Assessment is another area where there is a considerable impact of AI. Automated grading of assessments, such as multiple-choice questions, is already managed by assessment IT systems, but AI takes this one step further with the ability to grade more complex question formats. Natural language processing offers the ability to automatically assess short answers and essay responses, and to assess more than just accuracy (showing high correlation with human examiner marking) but also reasoning and argumentation (Schaye et al., 2025; Seneviratne & Manathunga, 2025). Apart from the grading, AI-based feedback promotes self-driven learning. By highlighting specific strengths and weaknesses from assessments, learners can take ownership of their progress and develop specific study strategies (Fazlollahi et al., 2022). This has been shown for MCQ assessments (Ch’en et al., 2025; Çiçek et al., 2025), short essays (Grévisse, 2024), OSCEs (Luordo et al., 2025), and clinical assessments (Schaye et al., 2022). This cycle of continuous practice, assessment, feedback, and improvement is a good fit with the principles of lifelong learning, which are fundamental in a rapidly changing medical environment.
In addition, Generative AI can easily produce formative and summative questions of various formats with proper prompting by an expert in the field who will then check the output. This helps greatly in reducing time spent by educators when designing assessments that are aligned with learning objectives. AI-generated questions can also be adjusted to various levels of difficulty and cognitive problems, which can be used for differentiated learning (Ahmed et al., 2025; Artsi et al., 2024; Law et al., 2025). These systems demonstrate that an AI–human pipeline can increase ease and efficiency of exam content development while maintaining expert validation. In addition, these question banks are capable of being updated continuously with data from learner performance to ensure assessments are relevant, fair and representative of current knowledge in medicine.
In the medical field, AI systems can analyze performance data from simulations or electronic health record (EHR) interactions. For example, algorithms can be used to measure diagnostic accuracy and time to intervention as well as adherence to clinical guidelines and therefore provide objective and immediate feedback (Kamel et al., 2024). This supports continuous education of clinicians in their respective fields.
When using AI for assessment, medical educators must consider some limitations and concerns which include validity and reliability risks (Grévisse, 2024; Schaye et al., 2022; Ten et al., 2020), potential factual errors and hallucinations (Zhang et al., 2024), limits in assessing nuanced communication (Rezayi et al., 2025), and transparency requirements (Quttainah et al., 2024).
3.6 Benefits of integrating IT and AI in medical education
The integration of IT and AI has revolutionized the field of medical education, unparalleled in the way students learn, practice their clinical skills, and are assessed. IT platforms have already extended the reach of teaching and learning content via digital libraries and through lectures, online classes, and learning applications in mobile platforms. This has been taken to the next level with the advent of AI, enabling individual paths for diverse learners. For example, adaptive systems adjust content difficulty for learners who need special support but offer advanced challenges for high achievers (Fontaine et al., 2019). This ensures equal access to learning, no matter how you look, how slowly, or the choice of how you prefer to learn (Ali et al., 2025).
Another benefit is the development of a greater ability to make informed decisions based on data for educators. IT is used to offer platforms for data collection; however, it is AI that will turn this data into an actionable form (Bojic et al., 2023; Buchanan et al., 2024). Through critical thinking, analysis of performance data, engagement data, and assessment data, it is possible for faculty to fine-tune teaching methods and reimagine curriculum. This provides the alignment of the instruction to the healthcare needs and continuous improvement in learning outcomes.
3.7 Ethical and pedagogical considerations
While the benefits of AI in medical education are significant and undeniable, there are important ethical and pedagogical issues to consider when adopting AI for medical education. Some concerns associated with algorithmic assessment are data privacy in the collection and analysis of learner performance data (Marcotte et al., 2025), the possibility of algorithmic bias in assessments (Li et al., 2025), and the potential for an over-reliance on AI at the cost of human judgment and empathy (Sauerbrei et al., 2023). Educators caution that although AI can be used to supplement learning, it needs to be used in a manner that supports rather than degrades humanistic elements of medical training practice. Training learners not only to use AI tools but also to evaluate the outputs critically from these tools is key to safe and ethical clinical decision-making (Ten et al., 2020).
3.8 Challenges and limitations of IT and AI in medical education and assessment
While integration of IT and AI holds great potential in enhancing the progression of medical education, its implementations are not without major challenges. These barriers have technical, institutional, ethical, and cultural components to their makeup and affect the uptake unevenly between regions and institutions.
The main challenge, therefore, is infrastructure and cost. The majority of AI platforms and simulation technologies require high-performance computing systems, strict internet connectivity, and software implementation updates (Roveta et al., 2025). Also, most institutions, especially in low- and middle-income countries, are not financially or technically well-placed to accommodate this type of innovation (Mergen et al., 2024). This inequity risks further digital divide with only better-resourced institutions able to utilize AI-enhanced teaching, learning and assessing. This is expected to heighten inequalities in medical training.
Another challenge is that of faculty preparation and training. As we transition from traditional teaching methods and move towards AI-assisted teaching and learning, educators will be required to have digital literacy and be flexible and adaptive. It is often the case that there are inconsistences in adoption due to faculty resistance and/or lack of preparedness (Khamis et al., 2025). Arguably, AI systems are only as useful as the settings in which they are appropriately used, and without adequate professional learning in the use of their tools, they may end up being underutilized or miss-utilized and consequently less useful.
4. Discussion
The present review situates Artificial Intelligence (AI) as an extension of the long digital transformation in medical education shifting emphasis from content delivery to experiential, adaptive learning. That shift is visible in the move from static e-learning toward immersive VR/AR and intelligent tutoring, where learners practice complex tasks safely, receive targeted feedback, and experience scenarios that would be rare or risky in clinical settings. These modalities support procedural memory, clinical reasoning, and team-based coordination, while enabling collaboration across institutions through multi-user simulations.
At the same time, practical implementation constraints remain. Substantial up-front investment, technical support requirements, and learning curves for both faculty and students are persistent barriers, especially in resource-constrained settings. These cost and capacity issues contribute to global disparities in access and uptake.
AI-enabled assessment is a particularly promising but a sensitive frontier. Generative systems can accelerate high-quality item development and allow continuous calibration of difficulty using learner data; video, speech, and interaction traces from simulations can provide real-time analytics of diagnostic accuracy, timeliness, and guideline adherence. However, these gains must be balanced against validity and reliability risks, transparency needs, and the possibility of subtle construct under-representation when assessing complex communication and professionalism. Governance and expert oversight are therefore non-negotiable.
AI platforms usually make use of enormous quantities of data, such as sensitive patient information. This makes organizations reluctant to use them due to fears about data ownership, algorithm bias, and other regulatory issues like educational and healthcare data regulations. If unaddressed, these issues may lead to loss of trust and insecurity among the learners and the institution.
Similarly, those who oppose AI argue that too much reliance on the technology can reduce the value of empathy, communication, and human judgment, which are essential in the field of medicine. The learners may also become overly dependent on the algorithmic feedback, which may adversely affect their ability to exercise independent clinical judgment.
A recurring theme is the imperative to preserve humanistic competencies. Over-reliance on automated feedback risks narrowing learners’ empathic engagement and independent clinical judgment. Programs should design AI as an augmentative partner (shaping formative feedback, remediation, and practice) while protecting bedside skills and reflective capacity through intentional pedagogy.
This review has also highlighted the need for AI literacy across undergraduate, postgraduate, and continuing education. Graduates must understand model behaviours, limitations, bias, and data stewardship to deploy tools safely in practice; curricula are beginning to incorporate these competencies, often alongside interdisciplinary standards for safe integration.
From a policy and leadership perspective, the synthesis of benefits and challenges is instructive. Benefits include expanded access, personalized pacing, data-informed curriculum improvement, and enhanced clinical preparedness through realistic simulations. Challenges cluster around infrastructure and cost, faculty readiness, privacy and ethics, risks of over-reliance, and unequal access; while underscoring that institutional readiness, governance, and equity strategies are essential to realize value at scale.
Methodologically, while this narrative review offers a panoramic, actionable map of technologies, it does not claim the inferential certainty of a registered systematic review with study-level bias appraisal and pooled estimates. Institutions should therefore pilot targeted use-cases, such as AI-supported simulation with structured debriefing, analytics-guided remediation, and faculty-supervised item generation, using robust evaluation designs and transparent reporting.
Looking forward, priorities include building interoperable infrastructure, faculty development at scale, and equitable financing models, particularly for low- and middle-income contexts, alongside a research agenda that links educational gains to clinical performance and patient-relevant outcomes. As VR/AR converges with AI, programs can leverage adaptive, multi-user, hyper-realistic environments, provided they embed ethics, data protection, and fairness by design. In short, AI should complement expert educators and authentic clinical experience, not replace them.
Table 1 highlights the key features of e-learning, simulation, virtual reality/augmented reality (VR/AR) and artificial intelligence (AI) for providing increased flexibility, personalization and assessment efficiency. However, barriers to widespread adoption include high costs, lack of infrastructure, unreadiness of faculties, privacy of data and bias, which necessitate the need for balance integration and continuous support by the institution.
Table 1. Summary of key technologies, benefits and implementation barriers for IT and AI in medical education and assessment.
| Key technology | Examples of use in medical education and assessment | Key benefits | Implementation barriers |
|---|---|---|---|
| E-learning Platforms | Online modules, LMS, virtual patient simulations, webinars. | Flexible, scalable access, standardized content delivery (Cook et al., 2008; Kononowicz et al., 2019; Papapanou et al., 2022). | Infrastructure and connectivity requirements, faculty development, engagement risks (Kassab et al., 2023; Regmi & Jones, 2020; Steinert et al., 2016). |
| Simulation Training | High-fidelity mannequins, task trainers, standardized patients. | Safe, repeatable practice, confidence and competence gains, effective debriefing (Bray et al., 2019; Cook et al., n.d.; Darnell et al., 2020; Maicher et al., 2023; Schaye et al., 2022). | High capital and maintenance costs, need for trained staff, limited realism (Alinier & Oriot, 2022; Jee et al., 2023; Maloney & Haines, 2016). |
| Virtual Reality/Augmented Reality | VR surgical/anatomy simulators, AR overlays for procedural guidance. | Immersive learning, practice of rare/high-risk cases, improved procedural performance (Bogomolova et al., 2020; Chien et al., 2022; Jung, 2022; Kaggwa et al., 2025; Kyaw et al., 2019; Mao et al., 2021; Wang et al., 2024b; Zhao et al., 2020). | Equipment costs, adaptation challenges, cybersickness, integration complexity (Li et al., 2024; Mergen et al., 2024; Raja & Al-Baghli, 2025). |
| AI and Intelligent Tutoring Systems | Adaptive quizzes, AI tutors for clinical reasoning, chatbots for Q&A. | Personalized learning at scale, targeted feedback, analytics-driven insights (Abe et al., 2025; Holderried et al., 2024; Kassab et al., 2023; Monteverde-Suárez et al., 2024). | Limited AI literacy, bias/validation concerns, data requirements (Car et al., 2025; Chan & Zary, 2019; Okamoto et al., 2025; Succi et al., 2025; Topol, 2023; Zhui et al., 2024). |
| AI enhances Virtual Reality/Augmented Reality | AI enhances VR/AR to make them more dynamic, realistic and reactive. | AI in VR/AR enhance realism, mimic broader simulations, provides feedback (Bray et al., 2019; Cook et al., n.d.; Darnell et al., 2020; Maicher et al., 2023; Schaye et al., 2022). | Equipment costs, adaptation challenges, cybersickness, integration complexity (Bray et al., 2019; Cook et al., n.d.; Darnell et al., 2020; Harmsen et al., 2024; Maicher et al., 2023; Schaye et al., 2022). |
| AI for Curriculum Development | AI curriculum mapping and alignment, analytics for gap analysis, generative AI for objectives/cases. | Data-informed planning, rapid updates, better alignment with competencies (Bojic et al., 2023; Buchanan et al., 2024; Harmsen et al., 2024; Kavvadias et al., 2020; Komenda et al., 2015; Waldvogel et al., 2025). | Low current adoption, faculty readiness, quality assurance, governance needs (Gordon et al., 2024; Khamis et al., 2025; Ten et al., 2020). |
| AI for Assessment | LLM-generated item banks, automated essay/SAQ grading, CV/NLP scoring for OSCE videos, analytics dashboards. | Faster item generation, consistent scalable grading, rapid feedback, targeted remediation (Ahmed et al., 2025; Artsi et al., 2024; Ch’en et al., 2025; Grévisse, 2024; Kamel et al., 2024; Law et al., 2025; Luordo et al., 2025; Schaye et al., 2025; Seneviratne & Manathunga, 2025; Çiçek et al., 2025). | Validity and reliability risks, potential factual errors, limits in assessing nuanced communication, transparency needs (Grévisse, 2024; Quttainah et al., 2024; Rezayi et al., 2025; Schaye et al., 2022; Ten et al., 2020; Zhang et al., 2024). |
Table 2 highlights the advantages of using IT and AI in medical education, such as improved accessibility, personalized learning, and realistic virtual clinical experiences. In addition, it allows for data-driven pedagogy, increased efficiency in educational delivery and constant improvement in the curriculum. Yet, the only way to have lasting success is to provide fair access, institutionalization, as well as proper training of the educators and learners.
Table 2. Five key benefits of IT and AI integration in medical education and assessment.
| Category | Description | Examples/Details |
|---|---|---|
| 1. Improved Accessibility and Inclusion | Enhances access to learning resources and supports diverse learner needs. | IT expands access through digital libraries, online lectures, and mobile apps (Cook et al., 2008; Kononowicz et al., 2019; Papapanou et al., 2022), AI enables adaptive learning paths that adjust content difficulty for individual learners while providing feedback (Fazlollahi et al., 2022; Rincón et al., 2025). |
| 2. Enhanced Clinical Preparedness | Provides realistic, risk-free clinical training experiences. | AI-driven simulations create dynamic, lifelike patient encounters to improve diagnostic accuracy, decision-making, and communication skills (Bray et al., 2019; Cook et al., n.d.; Darnell et al., 2020; Harmsen et al., 2024; Maicher et al., 2023; Prober & Heath, 2012; Schaye et al., 2022). |
| 3. Data-Driven Educational Decisions | Enables educators to personalize instruction using performance data. | AI analyzes assessment and engagement data to refine curricula and align teaching with healthcare needs (Abe et al., 2025; Fazlollahi et al., 2022; Holderried et al., 2024; Kassab et al., 2023; Monteverde-Suárez et al., 2024). |
| 4. Increased Learning Efficiency | Optimizes learning through intelligent automation and personalized pacing. | AI tailors content and feedback to the learner’s progress and comprehension level (Cook et al., n.d.; Schaye et al., 2022). |
| 5. Continuous Curriculum Improvement | Supports evidence-based enhancement of teaching methods. | Data insights help educators monitor and improve learning outcomes over time (Bojic et al., 2023; Buchanan et al., 2024; Harmsen et al., 2024; Kavvadias et al., 2020; Komenda et al., 2015; Waldvogel et al., 2025). |
Table 3 discusses the main challenges faced in using IT and AI in medical education: The major challenges involved in using IT and AI in medical education include high infrastructure and maintenance costs, lack of readiness in faculties and rise in data privacy concerns. Also, the overuse of technology may lead to the growing digitization of humanistic skills, while unequal access among institutions may lead to the widening digital gap in global medical education.
Table 3. Five key challenges of IT and AI in medical education and assessment.
| Category | Description | Examples/Details |
|---|---|---|
| 1. Infrastructure and Cost Barriers | High implementation and maintenance costs limit access. | AI systems require powerful computers, stable internet, and regular updates which are often unaffordable in low- and middle-income countries (Mergen et al., 2024; Roveta et al., 2025). |
| 2. Faculty Preparation and Training | Lack of digital literacy and resistance to change hinder adoption. | Educators may not be adequately trained to integrate AI tools effectively into teaching (Khamis et al., 2025; Ten et al., 2020). |
| 3. Data Privacy and Ethical Concerns | Risks related to data use and algorithmic bias. | Issues include patient data sensitivity, ownership, HIPAA/GDPR compliance, and maintaining trust (Chan & Zary, 2019; Li et al., 2025; Marcotte et al., 2025; Okamoto et al., 2025; Sauerbrei et al., 2023; Ten et al., 2020; Topol, 2023; Zhui et al., 2024). |
| 4. Overreliance on Technology | May undermine essential humanistic skills in medicine. | Learners might depend too heavily on AI feedback, weakening empathy and independent clinical reasoning (Okamoto et al., 2025; Sauerbrei et al., 2023; Topol, 2023). |
| 5. Unequal Access Across Regions | Creates a digital divide between institutions. | Well-resourced educational institutions benefit more, while the underfunded lag in implementation and innovation (Li et al., 2024; Mergen et al., 2024; Raja & Al-Baghli, 2025). |
5. Emerging trends and future directions of IT and AI in medical education and assessment
This latest wave of change is ushering in new paradigms in which IT and AI play an increasingly central role in medical education as their use continues to increase. These changes demonstrate not only the improvement of technologies, but also of the dynamic models of pedagogy, development of faculty and institutional priorities. The incorporation of AI literacy into the medical curriculum is one of the significant trends. The greater the medical schools regard the use of AI as a more than a teaching aid, a curriculum. The new courses introduce learners to the aspects of machine learning, algorithm decision making, and ethical issues in AI in healthcare. This indicates an increment in the addition of digital competencies to the curriculum, together with clinical knowledge. Increasingly, more institutions are establishing programs which integrate schooling in conventional medical sciences with education in data science, informatics, and digital ethics. This will help to ensure that graduates have critical reflection, along with the ability to responsibly apply AI tools to future clinical practice.
Looking into the future, IT is likely to play an even bigger role as VR and AR converge with AI to develop intelligent, adaptive and interactive learning environments. Current developments are merely an indication of a shift towards hyper-realistic, customized medical training settings that will be equipping medical students to take on more and more complex healthcare issues. Gamified patient simulators are a novel type of training equipment, which enable the learners to perform surgeries, engage in medical diagnostic decisions, and provide care to patients in the most lifelike looking virtual reality. These AI-enhanced simulations will allow educators to provide real-time, automated feedback, and be able to change the situation dynamically in response to the effectiveness of a learner.
There are also indications for automated and intelligent diagnostics in the future. AI is creating platforms for the assessment of not only factual knowledge, but also clinical reasoning, procedural, and communication skills. By analyzing patterns in learner responses, including speech, and even biometric data, these systems will provide more accurate, more focused and finer feedback that can be used to guide competency-based learning. In addition, the automated generation of valid, focused and quick assessments is becoming more and more feasible with generative AI.
Lastly, the IT/AI relationship in medical education and assessment is likely to occur in the context of more broad-based innovations in the healthcare sector, such as telemedicine, Internet of Things (IoT), robotics, and AI in radiology. As the world of clinical practice is constantly evolving along with technological advancement, medicine education will also be required to adjust along with the evolving environment to produce graduates who are well-equipped to operate in this new world of healthcare. Collectively, the trends are giving rise to a hybrid future of medical teaching, a teaching that integrates human-centered learning with the new intelligent and adaptive technologies that have the capacity to result in highly skilled, ethical, and visionary physicians.
6. Conclusions
Medical education is reflecting the changes which occur in the overall healthcare system. Since primitive forms of apprenticeship, Flexnerian reforms, and now digital-mediated enhancements, we are always seeking how to better prepare physicians to deal with the realities of clinical practice in the modern world. Possibly, the ultimate phase of the development is this adoption in the combined advances of Information Technology (IT) and Artificial Intelligence (AI). This is the latest and arguably the most impactful move in this trend.
In addition to widening access, crowdsourcing and encouraging collaboration, IT has accelerated the knowledge delivery process, and AI has transformed medical education due to the application of personalization, adaptive learning as well as high-fidelity simulation that recreates the rigor of real-life complexity. Combinations of these technologies can render medical education more inclusive, efficient and clinically relevant at the same time.
However, new issues related to infrastructure cost, faculty preparation, ethics and pedagogy exist. Addressing these issues is crucial to maximize the value of innovation in supporting medical education and assessment while not undermining the fundamental values that underlie the medical profession.
The trends in AI literacy, new instructional and immersive technologies, as well as the use of intelligent assessment, interaction with and feedback form digital healthcare systems are rising in the future, thus indicating a collaboration and mixed approaches to medical education. The model will seek to balance innovation in technology and humanistic learning. The future physician will be forced to be clinically competent, having a strong sense of ethics, and being flexible to rapid changes. Lastly, IT and AI are not to be perceived as the alternatives to the conventional approaches, but rather as powerful allies, which enrich the education and assessment processes with successful doses of modern medical training.
Declarations
Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work the authors used ChatGPT (OpenAI, GPT-5 Pro) to assist with language editing, section renumbering, and APA reference formatting. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Download
Export To
metrics
| Views | Downloads | |
|---|---|---|
| F1000Research | - | - |
|
PubMed Central
Data from PMC are received and updated monthly.
|
- | - |
Citations
CITE
how to cite this article
Shaban S and Magzoub ME. Artificial Intelligence in Medical Education and Assessment: The next step in the IT Revolution [version 1; peer review: awaiting peer review]. F1000Research 2025, 14:1360 (https://doi.org/10.12688/f1000research.173611.1)
NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.
track
receive updates on this article
Track an article to receive email alerts on any updates to this article.
Open Peer Review
Current Reviewer Status:
AWAITING PEER REVIEW
AWAITING PEER REVIEW
Key to Reviewer Statuses
VIEW
HIDE
ApprovedThe paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations
A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approvedFundamental flaws in the paper seriously undermine the findings and conclusions
Comments on this article Comments (0)