Doctors emphasize that the most valuable application of generative artificial intelligence (A.I.) in healthcare is its ability to alleviate the significant burden of documentation, which consumes hours of their daily work and contributes to burnout.
Dr. Matthew Hitchcock, a family physician in Chattanooga, Tennessee, benefits from an A.I. assistant that records patient visits on his smartphone and generates summaries for treatment plans and billing. After making some minor edits to the A.I.’s output, he completes his daily patient visit documentation in approximately 20 minutes. Previously, Dr. Hitchcock spent up to two hours typing medical notes after his children went to bed, but now he finds it “quite awesome” that those time-consuming tasks are a thing of the past.
While there is an inspiring vision of A.I. serving as a superintelligent partner to every doctor, providing care improvement suggestions, the current reality is that A.I.’s role in healthcare is more focused on administrative support and acting as a tireless scribe. Its primary function is to tackle the overwhelming volume of digital paperwork that physicians must produce for treatment, billing, and administrative purposes.
Experts, from leaders at major medical centers to family physicians, hold optimism that healthcare will benefit from the latest advancements in generative A.I., which possesses human-level fluency and can produce a wide range of outputs, from poetry to computer programs. However, they emphasize that medicine is not an open domain for experimentation, as A.I. algorithms for specific applications, such as scanning medical images for cancer detection, have undergone FDA approval. Chatbots are also being used by doctors to enhance communication with certain patients.
Physicians and medical researchers foresee slower adoption of generative A.I. in healthcare, particularly in the areas of diagnosis and treatment plans, due to regulatory uncertainties, concerns about patient safety, and the potential for litigation. Nonetheless, those who have utilized the technology affirm that its performance has significantly improved over the past year. The medical note software is designed in a way that allows doctors to verify the A.I.-generated summaries against the actual conversation with patients, instilling trust in the system.
The medical profession is plagued by high levels of burnout, leading many doctors and nurses to leave the field. Among the top complaints, especially for primary care physicians, is the time spent on documenting electronic health records, often referred to as “pajama time” when done after regular office hours. Generative A.I. holds promise as a potential solution to the physician workload crisis.
Various forms of documentation assistance, including speech recognition software and human transcribers, have been used by doctors for years. However, the latest A.I. technology goes beyond simple assistance by summarizing, organizing, and tagging conversations between doctors and patients.
Companies such as Abridge, Ambience Healthcare, Augmedix, Nuance (a Microsoft subsidiary), and Suki are developing this technology. For example, Abridge’s software is currently being used by over 1,000 physicians in the University of Pittsburgh medical system, enabling them to save hours in their day and engage more fully with patients.
Although generative A.I. holds significant potential, healthcare providers remain cautious due to concerns about the reliability and transparency of A.I.’s recommendations, as well as its tendency to occasionally produce fabrications or “hallucinations.” Ensuring patient safety and avoiding black-box decision-making are top priorities.
Dr. Hitchcock, who uses Abridge software, acknowledges the remarkable capabilities of A.I., but he refrains from relying on chatbots for diagnoses due to legal, regulatory, and practical considerations. Nevertheless, he is grateful that his evenings are no longer consumed by tedious digital documentation, which is a common requirement in the American healthcare industry. However, he does not see A.I. as a cure for the staffing shortages in healthcare and acknowledges that the field will still require human doctors for the foreseeable future.