Skip to content
After AI
Survives Durable: at least one of the four tests holds, and the evidence is direct.

Radiologists

The profession everyone predicted would vanish first, which instead got better at its hard half.

Protection from automation
Partly protected · our reading
−18.5%
Reading time cut for juniors
McKinsey 2025

This is the clearest case in the literature that exposure is not displacement. Expert opinion on both sides spent two decades predicting the end of radiology. What happened is in the trial data: less drafting, more reviewing and validating, and harder cases getting more human attention.McKinsey 2025

It is also the template for the whole site. PwC’s framework calls it professionalised — AI takes the routine work and leaves the expert work, so the role demands more skill and pays more, and those roles are growing twice as fast as the ones AI deskills.PwC 2026

What holds it up
  • The job is a bundle: drafting reports was never the valuable part
  • Diagnostic accountability sits with a named clinician
  • AI output needs verification, and verification is expert work
What threatens it
  • The drafting rung of the training ladder is disappearing
  • Radiology took 20 years of warnings and did not shrink — that is evidence, not comfort