ARI SHAPIRO, HOST:
With advances in artificial intelligence and robotics, more people are wondering whether their jobs are safe. And we're launching a new series to try to answer that question. We'll look at all types of work - low-paying gigs, lucrative professions, blue-collar, white-collar. And we'll need your help - more on that in a few minutes.
First, Lauren Silverman of member station KERA explores the job of a radiologist from inside the University of California, San Francisco.
LAUREN SILVERMAN, BYLINE: In health care, you could say radiologists have typically had a pretty sweet deal. They make on average around $400,000 a year, nearly double what a family doctor makes, and often have less grueling hours. But it became clear within minutes of entering the Department of Radiology. That golden path is no longer so secure.
MARC KOHLI: This is a reporter for NPR, and she's doing a story...
SILVERMAN: Hi, I'm Lauren.
KOHLI: ...On the future of our profession.
UNIDENTIFIED WOMAN: Oh great.
SILVERMAN: Inside a dimly lit reading room, young trainees click through black-and-white images of abdominal scans. In the last decade, the number of MRIs and CT scans has risen dramatically. Radiologists here assess 20 to 100 scans a day. Each scan can have thousands of images to review.
PHELPS KELLEY: Peribronchovascular pattern...
SILVERMAN: Every other minute, someone picks up a microphone to dictate into an electronic health record.
UNIDENTIFIED MAN: New paragraph - enlarged main pulmonary artery, comma...
SILVERMAN: For fourth year radiology fellow Phelps Kelley, the work is pretty predictable now, but...
KELLEY: The kind of biggest concern is that we could be replaced by machines.
SILVERMAN: Replaced by machines - that's because since radiology has gone digital, it's become an easy target for artificial intelligence, like machine learning, essentially complex problem-solving formulas. Dr. Bob Wachter is author of "The Digital Doctor."
BOB WACHTER: Radiology at its core is now a human being based on learning and his or her own experience looking at a collection of digital dots and a digital pattern and saying, that pattern looks like cancer or looks like tuberculosis or looks like pneumonia. Computers are awfully good at seeing patterns.
SILVERMAN: Just think about how Facebook's software can identify your face in a group photo or Google's can recognize a stop sign. Big tech companies are betting the same machine learning process, feeding a computer thousands of images could help it diagnose heart disease or stroke faster and cheaper than a human can.
KOHLI: I would say that there are many people who have a lot of angst about this.
SILVERMAN: UCSF radiologist Marc Kohli.
KOHLI: You can't walk through any of our meetings without hearing people talking about machine learning.
SILVERMAN: Both Kohli and his colleague Dr. John Mongan are researching ways to use artificial intelligence in radiology. Mongan says the people most fearful of AI understand the least about it. He compares the climate to that of the dot-com bubble.
JOHN MONGAN: People were very sure about the way things were going to go, you know? Webvan had billions of dollars and was going to put all the grocery stores out of business. There's still a Safeway half-mile down from my house. But at the same time, it wasn't all hype.
SILVERMAN: The reality is this. Dozens of companies, including IBM, Google and GE, are racing to develop formulas to diagnose diseases. Zebra Medical Vision, an Israeli company, already provides algorithms that help radiologists predict disease. Chief Medical Officer Eldad Elnekave says computers are great multitaskers. They can look for appendicitis while also checking for low bone density.
ELDAD ELNEKAVE: The radiologist can't make 30 diagnoses for every study, but the evidence is there. The information is in the pixels.
SILVERMAN: Still, UCSF's Mongan isn't worried about losing his job. In the short term, he's excited algorithms could help prioritize patients and make sure he doesn't miss something. Long-term, he says radiologists will spend less time looking at images and become...
MONGAN: The expert on those algorithms, the expert on how you apply them and how you interpret what comes out.
UNIDENTIFIED MAN: Pre-dominant of course-reticular opacities corresponding to fibrosis on the CT period...
SILVERMAN: Back in the reading room, radiology resident Phelps Kelley agrees. Machines won't replace radiologists, but radiologists who embrace AI will have a far more certain future. His analogy - Uber and the taxi business.
KELLEY: If the taxi industry had invested in ride-haling apps, then maybe they wouldn't be going out of business, and Uber wouldn't be taking them over. So if we can actually own this, then we can maybe benefit from it and not be wiped out by it.
SILVERMAN: For now, he offers what a computer can't - a diagnosis with a face-to-face explanation.
KELLEY: I can help you guys.
SILVERMAN: Lauren Silverman, NPR News.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.