Physician viewpoint: AI needs a 'Turing test' to determine medical feasibility

Though artificial intelligence has been proven in multiple studies to outperform human physicians at a variety of medical tasks, further tests are needed to determine whether this holds true in an actual clinical setting.

In an op-ed for STAT, Michael Joyner, MD, an anesthesiologist and physiologist at Rochester, Minn.-based Mayo Clinic, suggests a new kind of "Turing test" is needed before AI takes over complex medical tasks, many of which "lack fixed rules and stereotypical features" — two qualities typically necessary for an automated machine to do its job. (The original test was devised by British scientist Alan Turing in 1950 to measure a machine's ability to perform intricate tasks as well as humans.)

A suitable test, he writes, would be to enlist AI to create a successful weight loss plan for patients with severe obesity, since such a test "has the advantage of an easily measured outcome — all you need is a scale — and a condition that is potentially treatable by one or more interventions."

If a machine were able to devise an effective alternative to bariatric surgery, it "would clearly tip the scales and show the skeptics what medical AI can do," according to Dr. Joyner. "Or put more simply: It is time for medical artificial intelligence to go big or go home."

More articles on AI:
AdventHealth launches largest AI-enabled Mission Control command center in US
Google DeepMind co-founder placed on leave for alleged controversial projects
Tampa General opens high-tech clinical command center

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months