More exposure to AI would make med students less uneasy about radiology

During the fourth, fifth or sixth year of medical school, more than half of students across faculties in Brazil’s largest city believe AI is a threat to the radiology job market.

However, the trainees admit they’re going by surface perceptions. Some 64% acknowledge having insufficient knowledge about AI to confidently assess its impact on employment, and 32% would want more information on AI before deciding on radiology one way or the other.

The findings come from researchers at the Federal University of São Paulo in Brazil, the largest public university in the country, as published in the May/June edition of Radiologia Brasileira, which is published by the Brazilian College of Radiology and Diagnostic Imaging.

The team sent an anonymous online survey to 360 randomly chosen medical students at various São Paulo universities. They excluded students in the first three years of medical studies because literature shows early-year med students tend not to yet have formed strong preferences on specialties.

A total of 101 students in the target group completed the questionnaire. Along with the findings above, the authors report 74% (75 students) indicating they have never considered pursuing a career in radiology.

Among the 26 students who have, 18 ended up picking another path—11 because of AI and seven for other reasons.

Commenting on this finding in their discussion, the authors note that changes in career direction are common at the targeted stages of training even outside the context of AI, with up to 70% of students changing tracks for various reasons.

Still, they maintain, the findings of their survey “suggest that AI can have a negative impact on medical students’ choice of radiology as a specialty.”

“On the other hand,” the authors conclude, “there is evidence that indicates the existence of opportunities to minimize this trend by investing in education and quality information, clarifying realistically and objectively AI’s true role in the context of diagnostic imaging.”

The study is available in full for free (Google translation from Portuguese to English).