Radiologists in 2035: AI Will Expand, Concentrate, or Stratify the Profession
Three experts predict AI won't eliminate radiologists—but will reshape headcount and career security.
A new arXiv paper from Jan Beger, Amine Korchi, and Christoph A. Agten offers a structured disagreement on AI’s impact on diagnostic radiologists by 2035. Three subject-matter experts (two radiologists, one health tech professional with 20+ years in medical imaging IT) independently authored job descriptions using a shared template, each from a distinct vantage point: optimistic (AI expands the profession), trade-off (workforce economics concentrate roles), and professional stratification (sharply differentiated tiers). Across seven dimensions, all three scenarios agree on direction: AI will manage routine workloads, radiologists will carry accountability for AI output, and they will spend more time on complex cases and clinical collaboration. The disagreement is on magnitude—headcount, career security, and whether the profession broadens, shrinks into a well-compensated elite, or splits into tiers.
The clinical argument for optimism is real: radiologists offload mundane tasks and elevate their diagnostic and collaborative roles. But the economic argument for caution is equally valid: health systems may choose to concentrate or stratify the workforce to cut costs. The paper concludes that AI won't eliminate the diagnostic radiologist—but which future unfolds depends on choices health systems haven't made yet. For tech-savvy professionals, this highlights a critical inflection point: the technology is ready, but the organizational strategy is still up for grabs.
- All three scenarios agree: AI manages routine workload, radiologists focus on complex cases and clinical collaboration.
- Disagreement centers on headcount and career security—expansion, concentration, or stratification into differentiated tiers.
- Authors argue the clinical argument for optimism and economic argument for caution can both be true simultaneously.
Why It Matters
Healthcare leaders must choose which future to build—AI won't decide for them.