Cite as: Lu M, Ip V. Darwinian digitalization: modern evolutions in medical education. ASRA Pain Medicine News 2026;51. https://doi.org/10.52211/asra020126.012.

Introduction

The practice of medicine has historically been inextricably linked to the education of its practitioners. The tradition of Western medicine is generally considered to have originated in ancient Greece, where medical education was centered on the apprentice-practitioner relationship — a model in which a student would apprentice with a “physician-father” for several years.¹ The French Revolution led to the beginning of modern medical education when guilds were dissolved, causing medical education to fall purely under the purview of academia, which focused on public health rather than physician interests.² Perhaps one of the most significant contributors to modern medical education was Abraham Flexner, who advocated for university-based medical schools with full-time professors actively committed to research and recommended implementing minimum admission requirements alongside a rigorous, science-based curriculum grounded in the scientific method.^3,4 From the rationalization and standardization of medicine in the early 1900s to modern advancements, such as point-of-care ultrasound (POCUS), examining the evolution of medical education provides a framework to evaluate the integration of new technologies into curricula. This article reviews different types of educational delivery and examines how recent technological innovations, including online learning, artificial intelligence (AI), and POCUS, are shaping the future of undergraduate medical education.

Online Learning

In the modern era of medical education, advances continue to influence the future of medical education. One such advancement is the increasing use of remote learning or hybrid learning — a combination of remote and in-person instruction. This learning method came into sharp focus during the COVID-19 pandemic, as many schools transitioned to online classes due to social distancing and quarantine restrictions. A meta-analysis by He et al. found that synchronous distance education did not differ significantly from traditional education in learning outcomes but had higher student satisfaction ratings.⁵ Similarly, a systematic review by Vallée et al. of health learners from January 1990 to July 2019 found that hybrid learning led to better understanding of factual and conceptual course content compared to traditional learning alone for both subjective evaluations, such as learner self-reports, and objective evaluations, such as multiple-choice quizzes.⁶ Furthermore, a longitudinal study by Zhao et al. of Chinese medical students from three different universities developed custom scales that were administered to students in different learning modalities to evaluate satisfaction.⁷ They found that a hybrid approach resulted in the greatest student satisfaction, with solely traditional education coming in second and solely distance education coming in third.⁷ Zhao et al. speculated that the hybrid model allowed for more social interaction than purely online classes, which plays a significant role in student satisfaction, while also offering the convenience of online resources.⁷

These findings have implications for the future of medical education. Virtual and online education offer multiple benefits, including increased flexibility for students and educators, cost-effectiveness, a reduced carbon footprint, and broader dissemination of knowledge across institutions regardless of geography. Hybrid education maintains some interpersonal connection and networking, while hands-on practical skills mitigate the disadvantages of solely virtual platforms. These benefits, alongside increased student satisfaction, could shape the future of medical education.

Artificial Intelligence

With the advent of large language learning models (LLMs), AI has skyrocketed to the forefront of the public zeitgeist. Even the medical field has begun to integrate AI into clinical practice. Machine learning has been used to predict responses to chemotherapy, optimize drug dosing, and monitor public health trends.⁸ Natural language processing (NLP) has been used to create chatbots and virtual assistants that can schedule appointments and even triage patients.⁸ However, on the patient side, just as the proliferation of the internet led to easily searched medical advice, patients are already using LLMs like ChatGPT to get medical advice.⁹ The explosion in the popularity of AI has introduced some challenges. LLMs come with the risk of hallucinations, where the AI generates content that can seem plausible on the surface but may range from merely unverified information to completely falsified.¹⁰ These hallucinations can have wide-ranging consequences, from inaccurately triaging patients to eroding their trust in the medical system as a whole.¹⁰ Still, it is clear that AI is far from a passing trend and will likely shape the landscape of medicine going forward.

Although AI has become more widely accepted by the public and is infiltrating the administrative side of the medical field, undergraduate medical education (UME), or pre-residency medical school curricula, has been slow to adapt. The first study regarding the use of AI in medical education was published in 1992; however, there has been very little implementation of AI within all levels of medical education to date.^11,12 Although there has been some implementation in residency programs, such as radiology and cardiology, there is currently no official curriculum for AI in UME.^12,13There is a clear shortage of education, leaving medical students unprepared to address the challenges or to make use of the advantages that AI brings to clinical practice and academic medicine.

Virtual and online education offer multiple benefits, including increased flexibility for students and educators, cost-effectiveness, a reduced carbon footprint, and broader dissemination of knowledge across institutions regardless of geography.

However, there is some momentum to address this shortfall in UME, as it is a field that requires training for optimal utilization of AI and an understanding of its limitations. A 2021 literature review by Grunhut et al. found consensus that current physicians had insufficient knowledge of AI and that integrating AI into medical education was of great importance.¹⁴ A survey of 263 medical students at three German universities found that 71% of respondents indicated a need for AI to be included in UME.¹⁵

Yet, there is no universal consensus on the inclusion of AI in UME. It is an open secret that medical curricula are already constrained by insufficient time and resources. Thus, the benefits of introducing instruction on AI at the UME level are debated.¹⁶ At this level, medical students are still developing the fundamental knowledge base, skills, and professional competencies of a physician.¹⁶ At the UME level of training, students have not yet achieved the proficiency required to utilize AI to augment healthcare practices properly. As such, some argue that instruction on AI is fruitless without the requisite insight into healthcare practice to implement AI properly.¹⁶ In fact, it may be detrimental – increased reliance on AI might hamper the development of clinical reasoning in medical students.¹⁶ Since the foundation of the basic sciences, such as physiology, anatomy, and pharmacology, is essential for developing skills for clinical diagnosis, it may be more important to create a solid foundation rather than complicate their development with the introduction of AI.¹⁶ Nonetheless, the importance of understanding and learning the proper integration of AI in clinical and academic medicine is paramount.

However, there are challenges to integrating AI into the medical curriculum due to a lack of knowledge within the medical community.^12,14 It is extremely difficult to develop a curriculum when expertise is lacking in the medical field. This is further compounded by the scarcity of research on integrating AI into medical education. There is a distinct lack of curricular frameworks for AI, which means that any attempt to integrate AI will have to be fully built from the ground up.¹² This fundamentally increases the amount of resources required to construct a curriculum, an undertaking compounded by a lack of experts. There is a questionable risk-benefit ratio regarding resource utilization and cost, given the actual utility of AI in UME.

Simulation Training and AI

Another effect of AI on medical education has been the advancement of simulation training. Training medical students has always carried an essential paradox: Trainees must receive practical experience to develop into competent physicians in clinical practice, but their inexperience during training can compromise patient safety and care.¹⁷Simulation-based training (SBT) emerged as a means for trainees to refine their skills in a safe, controlled environment, thereby reducing patient risk and the inherent variability of a clinical setting.¹⁷ As a consequence, SBT is inherently unable to truly simulate a real clinical setting, which can undermine the transferability of skills developed in a simulator to the real world.¹⁷ AI has been shown to enhance the realism and dynamism of SBT through dynamic patient responses and the ability to holistically adjust the scenario based on learners' actions, particularly with advances in virtual and augmented reality.¹⁷ However, a study by Kollerup et al. found that residents using a transvaginal ultrasound simulation were frustrated by the system’s limited feedback.¹⁸ Residents were often unable to understand the error they had made, how to resolve it, or even what the correct procedure was. Many were, therefore, forced to make errors deliberately to proceed with the simulation.¹⁸ Hence, Kollerup et al. recommend that AI should be integrated to provide more flexible, live, and continuous feedback to support resident learning.¹⁸ Another aspect of SBT has been standardized patients (SP). Use of SPs has been invaluable for medical training; however, a major limitation is the variability in human performance and the inherent inability to fully control all aspects of the simulated interaction.¹⁹ Simulations that use AI can overcome this issue. With advances in NLP, AI simulations can maintain realistic dialogue and interactions with trainees while remaining highly consistent, allowing students to practice without substantial variation across sessions repeatedly.¹⁹

Nevertheless, AI cannot replace the entirety of medical training. A report by Aggarwal et al. evaluated simulations and their ability to support the development of the CanMEDS roles.²⁰ In Canada, the CanMED is a competency-based framework the Royal College of Physicians and Surgeons of Canada developed to define the essential skills for medical professionals.²¹ To provide high-quality patient care, physicians should master the roles. This continuum, from medical students to consultants, guides education, training, and assessment in medicine. They found that while simulations were effective to promote the roles of medical expert, collaborator, and communicator, there was limited use or evidence for scholar, professional, manager, or health advocate.²⁰ Indeed, recommendations for AI feedback and training are designed to supplement human feedback rather than supplant it.¹⁸ Furthermore, simulators are incredibly expensive and resource-heavy.¹⁷ Although introducing AI into simulators will increase the fidelity of said simulators, it will also vastly increase their complexity. Complex, high-fidelity simulators already require incredible capital, and introducing AI will only compound these resource requirements.¹⁷ It also must be considered that much of the benefit of simulators to date comes from lower fidelity simulators, which calls into question the need to introduce AI.²⁰ The use of AI can also lead to unintended outcomes, which must be resolved. Although a study by Fazlollahi et al. found that medical students who used an AI-augmented simulation curriculum caused 55% less tissue damage, they also found that their operations were less efficient, with tumour tissue removal 29% slower.²⁰

Although the general public has increasingly accepted AI, AI and medical training are still in a very nascent stage. In 2018, the American Medical Association advocated for increased research about AI instruction in medicine. Though progress is slow, there has been an increased interest in AI.²² Just last year, a paper from Fletcher described the scaling up of a clinical AI fellowship at the National Health Service , demonstrating a cogent interest in developing new AI training.²³ To best prepare future physicians for the complexities of AI, in both education and training, more research must be completed on creating curricular frameworks, providing background education on AI, and implementing the integration of AI into UME curricula.^14,24

Point of Care Ultrasound

The past few years have seen a marked increase in the use of POCUS. As technology has advanced, previously unwieldy and bulky ultrasound (US) machines have been miniaturized, and costs have decreased dramatically, leading to widespread adoption of POCUS.²⁵ Furthermore, the effectiveness of POCUS cannot be understated – aside from increasingly being used in a variety of specialties, first-year medical students were more accurate at identifying cardiac abnormalities with 18 hours of POCUS training than cardiologists were with standard methods, and trained medical students were able to more accurately measure liver size with POCUS than certified internists were with standard methods.^25–27 As such, there has been more interest in integrating POCUS into UME.

Although the Liaison Committee on Medical Education currently has no mandatory UME training requirement, a 2012 survey of 134 curricular administrators at United States medical schools found that 62% of these schools had integrated US training into their curricula.^28,29 Of these schools, the majority of US instruction occurred in the third year, when it focused on obtaining and interpreting US scans.²⁹ Those who taught US in the first 2 years indicated that it mainly served as a tool for basic science or introductory topics.²⁹ Interestingly, though 79% of respondents supported the integration of the US into the UME curriculum, only 14% indicated that US education was a priority.²⁹ Similarly, a 2014 survey of 13 medical schools in Canada found that only six had integrated POCUS into their UME curriculum.³⁰ All of the schools had practical instruction alongside an instructor where students operated a probe on live models or in simulations, but 67% also included a theoretical format.^30,1 US education took place in a combination of anatomy labs, medical simulation centers, classrooms, and hospitals, with instructors who were non-radiologists trained in the US.³⁰ Interestingly, a majority (77%) of vice-deans of these schools indicated that POCUS should be taught during UME; however, many cited the lack of suitable US machines and infrastructure as barriers to integrating POCUS.³⁰

The Canadian Ultrasound Consensus for Undergraduate Medical Education Group, established in March 2018, convened a panel of 21 US experts from 15 Canadian medical schools to identify curricular elements to be integrated into UME education.³¹ Furthermore, the Seguin Canadian POCUS Education Conference formed a panel comprised of medical students and physicians from 14 different medical schools that generated a series of 14 recommendations for POCUS implementation.³² While it is clear that there is interest from medical students in implementing POCUS in UME, the already overwhelming medical curricula remain a barrier to POCUS integration.

The Brody School of Medicine (SOM) at East Carolina University in Greenville, NC, successfully incorporated US training with minimal faculty and curricular time.^29,30,32,33 Rathbun et al. describe how US training was first incorporated as a half-day emergency medicine (EM) clerkship, then as a 4-week fourth-year medical student elective, and finally as a series of voluntary 1-hour hands-on sessions corresponding to appropriate anatomy and physiology courses.³³ Overall, students from all three phases overwhelmingly found that US education was helpful, appropriate for their level of education, and wished that they had been exposed to the US earlier in their medical training or that it was incorporated into their medical curriculum.³³ Rathbun et al. were able to reduce resource usage by integrating much of the didactic teaching into online modules, which they found required much less faculty time once the resources had been created.³³ Furthermore, the optional extracurricular sessions made use of EM residents and peer tutors to work with the first-year students while a single EM faculty member circulated as an instructor.³³ While this peer-teaching model allowed for greater scalability with minimal faculty time, the relative inexperience of second-year medical students could affect the quality of instruction, which could be avoided for future implementation of this model.³³ Though the authors do note that this model may not be applicable at every school, the success of the Brody SOM at integrating POCUS education into UME provides a strong stepping stone that can be adjusted to the needs and barriers that different schools face.³³

Conclusion

Medical education has faced numerous barriers, from the early descriptions of apprenticeships to the modern two-tiered preclinical and clinical education widely used around the world. As medical technology and practice continue to evolve, faculty must incorporate more content into an already full medical curriculum. Nonetheless, there is increasing pressure to integrate emerging technologies into medical education to adequately prepare physicians for a changing medical landscape.

References

1. Fulton JF. History of Medical Education. Med. J 1953;2(4834):457-61. https://doi.org/10.1136/bmj.2.4834.457
2. Custers EJFM, Cate OT. The history of medical education in Europe and the United States, with respect to time and proficiency. Med 2018;93:S49-S54. https://doi.org/10.1097/ACM.0000000000002079
3. Buja LM. Medical education today: all that glitters is not gold. BMC Med. Educ 2019;19(1):110. https://doi.org/10.1186/s12909-019-1535-9
4. Barzansky B. Abraham Flexner and the era of medical education reform. Med 2010;85(9 Suppl):S19-25. https://doi.org/10.1097/ACM.0b013e3181f12bd1
5. He L, Yang N, Xu L, et al. Synchronous distance education vs traditional education for health science students: a systematic review and meta-analysis. Educ 2021;55(3):293-308. https://doi.org/10.1111/medu.14364
6. Vallée A, Blacher J, Cariou A, et al. Blended learning compared to traditional learning in medical education: systematic review and meta-analysis. Med. Internet Res 2020;22(8):e16504. https://doi.org/10.2196/16504
7. Zhao Y, Sun T, Zhang X, et al. The evolution of medical education in the era of Covid-19 and beyond: a longitudinal study. BMC Med. Educ 2024;24(1):1289. https://doi.org/10.1186/s12909-024-06271-8
8. Alowais SA, Alghamdi SS, Alsuhebany N, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med. Educ 2023;23(1):689. https://doi.org/10.1186/s12909-023-04698-z
9. Cooper A, Rodman A. AI and medical education — a 21st-century pandora’s box. Engl. J. Med2023;389(5):385-7. https://doi.org/10.1056/NEJMp2304993
10. Qin H, Tong Y. Opportunities and challenges for large language models in primary health care. Prim. Care Community Health 2025:16:21501319241312571. https://doi.org/10.1177/21501319241312571
11. Gordon M et al. A scoping review of artificial intelligence in medical education: BEME Guide No. 84. Teach 2024;46(4):446-70. https://doi.org/10.1080/0142159X.2024.2314198
12. Tolentino R, Baradaran A, Gore G, et al. Curriculum frameworks and educational programs in AI for medical students, residents, and practicing physicians: scoping review. JMIR Med. Educ 2024:10:e54793. https://doi.org/10.2196/54793
13. Singla R, Pupic N, Ghaffarizadeh SA,et al. Developing a Canadian artificial intelligence medical curriculum using a Delphi study. Npj Digit. Med 2024;7(1):323. https://doi.org/10.1038/s41746-024-01307-1
14. Grunhut J, Wyatt AT, Marques O. Educating future physicians in artificial intelligence (AI): an integrative review and proposed changes. Med. Educ. Curric Dev 2021:8:23821205211036836. https://doi.org/10.1177/23821205211036836
15. Pinto dos Santos D, Giese D, Brodehl S, et al. Medical students’ attitude towards artificial intelligence: a multicentre survey. Radiol 2019;29(4):1640-6. https://doi.org/10.1007/s00330-018-5601-1
16. Ngo B, Nguyen D, vanSonnenberg E. The cases for and against artificial intelligence in the medical school curriculum. Artif. Intell 2022;4(5):e220074. https://doi.org/10.1148/ryai.220074
17. Elendu C,Amaechi DC, Okatta AU, et al. The impact of simulation-based training in medical education: a review. Medicine (Baltimore) 2024;103(27):e38813. https://doi.org/10.1097/MD.0000000000038813
18. Kollerup NK, Johansen SS, Tolsgaard MG, et al. Clinical needs and preferences for AI-based explanations in clinical simulation training. Inf. Technol 2025;44(5):954–74. https://doi.org/https://doi.org/10.1080/0144929X.2024.2334852
19. Zidoun Y, Mardi AE. Artificial intelligence (AI)-based simulators versus simulated patients in undergraduate programs: a protocol for a randomized controlled trial. BMC Med. Educ 2024;24(1):1260. https://doi.org/10.1186/s12909-024-06236-x
20. Aggarwal R, Mytton OT, Derbrew M, et al. Training and simulation for patient safety. BMJ Qual. Saf 2010:19 Suppl 2:i34-43. https://doi.org/10.1136/qshc.2009.038562
21. Frank JR, Snell L, Sherbino J. CanMEDS 2015 Physician Competency Framework. Ottawa, Canada: Royal College of Physicians and Surgeons of Canada, 2015.
22. Augmented intelligence in health care. American Medical Association. https://www.ama-assn.org/system/files/2019-01/augmented-intelligence-policy-report.pdf . Published June 2018. Accessed Month Day, Year.
23. Fletcher B. Scaling up a clinical AI fellowship. Health Technol. Inform 2024;315:637-8. https://doi.org/10.3233/SHTI240254
24. Solomon SD, Saldana F. Point-of-care ultrasound in medical education - stop listening and look. Engl. J. Med 2014;370(12):1083-5. https://doi.org/10.1056/NEJMp1311944
25. Kobal SL et al. Comparison of effectiveness of hand-carried ultrasound to bedside cardiovascular physical examination. J. Cardiol 2005;96(7):1002-6. https://doi.org/10.1016/j.amjcard.2005.05.060
26. Mouratev G, Howe D, Hoppmann R, et al. Teaching medical students ultrasound to measure liver size: comparison with experienced clinicians using physical examination alone. Learn. Med 2013;25(1):84-8. https://doi.org/10.1080/10401334.2012.741535
27. Gilbertson EA, Hatton ND, Ryan JJ. Point of care ultrasound: the next evolution of medical education. Transl. Med 2020;8(14):846. https://doi.org/10.21037/atm.2020.04.41
28. Bahner DP, Goldman E, Way D, et al. The state of ultrasound education in U.S. medical schools: results of a national survey. Med. J. Assoc. Am. Med. Coll 2014;89(12):1681-6. https://doi.org/10.1097/ACM.0000000000000414
29. Steinmetz P, Dobrescu O, Oleskevich S, et al. Bedside ultrasound education in Canadian medical schools: a national survey. Med. Educ J 2016;7(1):e78-86.
30. Ma IWY, Steinmetz P, Weerdenburg K, et al. The Canadian medical student ultrasound curriculum. Ultrasound Med 2020;39(7):1279-87. https://doi.org/10.1002/jum.15218
31. Weill S, Picard DA, Kim DJ, et al. Recommendations for POCUS curriculum in Canadian undergraduate medical education: consensus from the inaugural Seguin Canadian POCUS Education Conference. POCUS J2023;8(1):13-18.. https://doi.org/10.24908/pocus.v8i1.16153
32. Rathbun KM et al. Incorporating ultrasound training into undergraduate medical education in a faculty-limited setting. BMC Med. Educ 2023;23(1):263. https://doi.org/10.1186/s12909-023-04227-y