The Chatbot Will See You Now: How AI Is Changing Medicine
Posted by Richard on September 7, 2023
Artificial intelligence is unlikely to replace doctors or nurses — at least, not anytime soon. But in an industry that still runs mostly on time-consuming manual processes, AI breakthroughs are set to permanently change how clinicians do their jobs and how patients access care.
Already, generative AI is in use in some clinics and hospitals as a sort of smart digital assistant to automate routine tasks like clinical documentation, according to the New York Times. One Tennessee-based family physician estimates that he spends 20 minutes finishing documentation at the end of each day with the help of AI software that records and summarizes visits, compared to the two hours of extra work he tackled each night with manual documentation.
AI-powered chatbots that collect and analyze information from patients are also growing in popularity, according to Forbes. Digital health startup K Health’s technology converses with patients about their symptoms, checks that information against millions of other patients, delivers a summary with a list of possible diagnoses to a human clinician for official diagnosis and treatment, and documents the entire thing in the patient’s record. According to the company’s founder, more than 3.1 million patients have “seen” a doctor or nurse through the app so far for services like primary care, urgent care, some pediatric services, and chronic disease management.
While AI might provide a lifeline to overworked doctors and make it easy to access basic health information, some experts urge caution about using it as a diagnostic tool, according to Scientific American. Epidemiologist Andrew Beam of Harvard University is concerned about AI’s susceptibility to misinformation, while other physicians are concerned that it could be misused in medical education.
Chatbots are also prone to hallucinate new information, sometimes complete with references that don’t exist. In May 2023, one New York lawyer gained national attention after he submitted ChatGPT-generated research that included fake cases and bogus citations.