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.

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).

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).

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).

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).

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.

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).

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.

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.