Medical Student Perspectives on the Impact of Artificial Intelligence on the Practice of Medicine
Introduction
Interest and research in artificial Intelligence (AI) has increased dramatically over the last decade1 with billions of dollars projected to be spent towards Deep Learning (DL), in particular.2 As DL has become increasingly applied to medical image analysis, prominent computer scientists have made ominous predictions about the future of radiology, ranging from DL algorithms being able to outperform radiologists in image interpretation4 to calls to “stop training radiologists now.”3 Such sentiments have unsurprisingly been covered heavily by the lay media, both within medical and technology nches4 and more broadly in mainstream outlets.5
Such negative assessments of the future of radiology in relation to AI and DL might be expected to cause medical students to have an unfavorable view of the future of radiology as a specialty to enter. In fact, studies performed in Canada and Europe have shown that AI may be significantly impacting the way that medical students approach their choice of specialty in a negative manner.6,7 In response, a recent editorial has proposed that radiology educators should proactively address these concerns and view AI and DL as an opportunity recruit medical students to a field that will interface intimately and collaboratively with these exciting technologies.
Although the perceptions of medical students towards radiology in relation to AI have been evaluated in both Canada and Europe,6,7 no such study has been performed in the United States. The purpose of this study was thus to evaluate the perceptions of medical students about radiology and other medical specialties in relation to AI.
Section snippets
Survey Administration & Participants
We conducted a survey of medical students in the United States from November 2017 to April of 2018 titled “Artificial Intelligence and the Future of Medicine,” with the goal of evaluating perceptions towards radiology and other medical specialties in relation to AI. We sent our survey in the form of a browser-based, anonymous questionnaire (SurveyMonkey.com, North America) to 32 radiology interest groups at medical schools across the United States chosen to provide a representative mix of
Survey Respondent Summary
There was a total of 156 responses with a very high survey completion rate of 95.4% (of all students who started the survey). There was representation from students from each year (M1: 25.8%, M2: 27.1%, M3:17.4%, M4:29.7%) of medical school (Fig 3).
The effect of AI on the Practice of Medicine During their Career
Over 75% of students revealed that they believed AI would have a moderate-to-major effect on medicine during their careers (Fig 4). A small minority (1.9%) reported that AI would play no part in the future of medicine. On sub-analysis by medical
Discussion
Due to anecdotal reports of medical student concerns regarding AI's influence on the future practice of medicine, as well as sentiments from both medical professionals and the lay media predicting deleterious effects of AI on the viability of diagnostic radiology, we investigated whether there could be a negative impact of AI on medical students’ perceptions of the field. Although medical student perceptions of AI's impact on radiology have been studied in Canada and Europe,6,7 no similar study
Conclusion
The continued misperception of AI as a replacement for radiologists is much more of a threat than the development and integration of AI into the practice of diagnostic imaging, as our study suggests that it may erode medical student enthusiasm for the field. This has to potential to reduce incoming talent into the field and ultimately result in a “brain drain.” Radiology has always been at the forefront of technology utilization and education in medicine – MRI, CT, and picture archiving and
Appendix
List of Contacted Radiology Interest Groups
Boston University School of Medicine
University of Pennsylvania Perelman School of Medicine
University of Alabama Birmingham School of Medicine
University of Washington School of Medicine
Howard University College of Medicine
University of Arkansas for Medical Sciences College of Medicine
University of Michigan Medical School
Wayne State University School of Medicine
University of Arizona College of Medicine
Wake Forest School of Medicine
University of Maryland
References (8)
- et al.
Influence of artificial intelligence on Canadian medical students’ preference for radiology specialty: A national survey study
Acad Radiol
(2019) - et al.
The day when computers read between lines
Jpn J Radiol
(2019) - Worldwide Spending on Cognitive and Artificial Intelligence Systems Will Grow to $19.1 Billion in2018, According to New...
- Mukherjee SAI, Versus MD. Published online March 27, 2017. Available at:...
Cited by (53)
Systematic Review of Radiology Residency Artificial Intelligence Curricula: Preparing Future Radiologists for the Artificial Intelligence Era
2023, Journal of the American College of RadiologyRadiology and the Media
2023, Academic Radiology
Conflict of Interest: None.