Nearly 10 years after dire predictions that artificial intelligence (AI) would wipe out radiology, the field is facing a shortage, rising demand and record-high pay.
Radiologist Shortage Drives Salaries Higher
In 2016, Geoffrey Hinton, the “Godfather of AI," said it was "completely obvious" machines would outperform radiologists within 5 to 10 years, as reported by Fortune on Sunday.
He even compared the profession to "the coyote that's already over the edge of the cliff."
A decade later, the opposite has unfolded. The number of practicing radiologists in the U.S. has grown modestly, but demand has surged far faster.
There were more than 4,000 open radiologist roles as of early 2025, with positions taking months to fill. Average annual salaries have climbed to about $571,000, according to industry data.
"We actually have a huge shortage of radiologists. So the exact opposite of this prediction has happened," said Christoph Herpfer, a University of Virginia economist who studies physician labor markets.
Experts say earlier forecasts misunderstood the scope of the job.
While AI has improved at reading medical images, radiologists do far more, including consulting with physicians, guiding treatment decisions and performing procedures.
AI Reshapes Jobs, Not Replaces Them, Say Tech Leaders
Earlier, Nvidia Corp. (NASDAQ:NVDA) CEO Jensen Huang said AI was transforming industries and creating new jobs as part of a broader industrial shift, even as it automated certain tasks.
He said engineers remained essential to ensuring AI systems were safe, useful, and beneficial, arguing the technology was expanding opportunity rather than eliminating work.
Amazon.com Inc. (NASDAQ:AMZN) introduced an AI hiring tool called "Connect Talent" to automate parts of large-scale recruitment, including screening and interviewing candidates for high-volume roles.
The company said the system was designed to improve efficiency while still keeping candidates informed when AI was involved.
Box Inc. (NYSE:BOX) CEO Aaron Levie said fears of widespread job loss from AI were overstated, noting that workers often underestimate the human effort required to make AI outputs usable.
He pointed to tasks like verifying data, adding context, and integrating results into real workflows as critical steps that AI alone could not handle.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Photo courtesy: Shutterstock
Login to comment