Artificial intelligence has the potential to influence physicians' decision-making or flag potential serious health concerns for patients, but becoming overly reliant on the technology is risky, according to a viewpoint from Jonathan Leibowitz, MD, published in The Wall Street Journal Nov. 12.
Four things to know about the potential and pitfalls of AI in healthcare:
- Dr. Leibowitz is a primary care physician in New York City. Dr. Leibowitz is also board certified in internal medicine and nephrology. AI can be a great asset in reading pathology slides, but it's incapable of reading people, Dr. Leibowitz said. It is risky to train new physicians to depend on algorithms in their medical school training.
- Dr. Leibowitz said it's equivalent to training a new driver to get around just by using GPS navigation. This driver would be incapable of going around the block on their own, a possibility that Dr. Leibowitz said is "not so smart."
- AI is becoming increasingly popular to use in hospitals to automate tricky or tedious tasks, such as chatbots for patient triaging. New York City-based Mount Sinai, Cleveland Clinic and Edison, N.J.-based Hackensack Meridian Health have all recently deployed AI tools to assist physicians in diagnosing patients.
- Some experts have expressed concern that AI has the capability of compounding existing biases, while advocates for AI said it has the potential to reduce clinician burnout, accelerate drug discovery and more.