Impact of artificial intelligence on US medical students' choice of radiology

Abstract

PurposeInternational student surveys have shown significant anxiety about pursuing radiology as a career due to artificial intelligence (AI). For a counterpart study in the US, we examined the impact of AI on US medical students' choice of radiology as a career, and how such impact is influenced by students' opinions on and exposures to AI and radiology. MethodsStudents across 32 US medical schools participated in an anonymous online survey. The respondents' radiology ranking with and without AI were compared. Among those considering radiology within their top 3 choices, change in radiology ranking due to AI was statistically examined for association with baseline characteristics, subjective opinions, and prior exposures. ResultsAI significantly lowered students' preference for ranking radiology (P < .001). One-sixth of students who would have chosen radiology as the first choice did not do so because of AI, and approximately half of those considering radiology within their top 3 choices remained concerned about AI. Ranking radiology lower due to AI was associated with greater concerns about AI (P < .001), less perceived understanding of radiology (P = .02), predicting a decrease in job opportunities (P < .001), and exposure to AI through medical students/family (P = .03) as well as through radiology attendings and residents (P = .03). Education on AI during radiology rotations, followed by pre-clinical lectures, was the most preferred way to learn about AI. ConclusionAI has a significantly negative impact on US medical students' choice of radiology as a career, a phenomenon influenced by both individual concerns and exposure to AI from the medical community.