Dr. Watson: How IBM’s supercomputer could improve health care
By Martin Ford,
Watson, the IBM supercomputer that defeated the world’s best “Jeopardy!” playersthis year, has found a job in medicine. It won’t be consulting with patients, but a version of the game-show champion could appear in examination rooms, offering assistance to flesh-and-blood physicians. But how soon might you see Dr. Watson? And could Dr. Watson be better than your doctor?
I’ve worked in software development for more than 25 years — never for IBM — and was amazed by Watson’s ability to understand language, solve problems and present answers in the form of a question as Alex Trebek coolly looked on. While I don’t think most doctors need to worry about their jobs anytime soon, Watson-esque technology offers a powerful diagnostic tool that could bring dramatic benefits to health care. The prospect of a robotic caregiver might not seem comforting, but Watson’s first appointment will be a watershed moment. This is an app way beyond anything on an iPhone.
Machine learning — the technology that makes Watson, well, Watson — lets software analyze historical data and generate rules that inform decision-making, even when tasks significantly vary. Watson could churn through millions of case histories to learn what diagnosis is likely to be correct and what treatment would be the most effective. The system could almost instantly process medical textbooks, electronic medical records and the latest published research, illuminating obscure links among studies in seemingly unrelated specialties. Watson could someday be a standard diagnostic tool. Its ability to make sense of a universe of data would be far beyond that of any person or team of experienced physicians.
Watson could dazzle in the exam room. It could lead to greater consistency of care, giving every patient an exhaustive review of treatment options. It’s a good bet that this smart machine will evolve into a genuine Dr. Watson: a true diagnostic partner for physicians, offering not just a reference but a viable second opinion.
This seems certain to help contain health-care costs. For one thing, it could result in fewer errors and better patient outcomes, protecting physicians from malpractice claims. For example, in 2008, the Supreme Court ruled in Riegel v. Medtronic that medical-device manufacturers facing lawsuits aren’t necessarily liable if the FDA has approved their products. If that legal reasoning is someday extended to include diagnostic systems and their users, doctors relying on Watson might not need as much insurance, addressing their complaints about the costs of practicing medicine and passing on the savings to patients, insurers and Medicare.
Watson could also dramatically change the way medical services are delivered. Once a machine with diagnostic expertise, no need for sleep and a proven track record can provide quality care, will it really be necessary for a physician to see every patient? We might want to create a new category of medical professional: people trained to examine patients and convey their symptoms to a system such as Watson. These lower-paid practitioners could help manage the growing number of patients with chronic conditions such as obesity and diabetes, practicing medicine without going to medical school.
Physicians and the American Medical Association would resist the influx of these less-educated colleagues. But let’s face it: Most medical school graduates have little interest in entering family practice and even less in serving rural areas. The Association of American Medical Colleges predicts a shortage of up to 150,000 doctors within the next 15 years as older practitioners retire, the Obama administration’s health-care law insures millions of new patients and an aging population requires ever more care. There will be a shortage of primary-care physicians as medical school graduates burdened by onerous student debt enter more-lucrative specialties.
And Watson will find applications beyond medicine. While doctors wouldn’t be completely displaced by machines — Watson wouldn’t have a medical license, and patients might prefer that a human being deliver news of a poor prognosis — the same cannot be said of corporate drones. As artificial intelligence becomes more capable, people engaged in any kind of routine work will eventually become redundant. When machines can learn, many jobs we now consider safe from automation — such as insurance underwriters, financial analysts and even lawyers — will be threatened.
What would this do to our economy? Beyond medicine, Watson’s debut isn’t entirely good news. The Industrial Revolution and the Internet boom proved that the economy can adapt to advancing technology and create new jobs requiring new skills. But today’s information technology is taking on cognitive decision-making — the kinds of analytical tasks that we have never outsourced to machines. Future Watsons may erode the high-skill, high-wage jobs essential to the nation’s prosperity.
“There’s no shame in losing to silicon,” “Jeopardy!” champ Ken Jennings wrote after being bested by Watson. But losing a game show is one thing. Whether you’re a doctor, a lawyer or a journalist, losing a job is another.
Martin Ford is the author of “The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future.” He blogs at econfuture.wordpress.com.
Read “With IBM’s Watson on ‘Jeopardy!’, how trivial is trivia?” by Alexandra Petri.