If you are in charge of healthcare and want to help your clinicians build a ChatGPT tool, you should look at the latest study coming out of Boston. Researchers at Mass General Brigham have found that the ?large language model (LLM) AI chatbot is 72% accurate at making clinical decisions across all medical specialties and stages of care. The tool is also 77% accurate at making a final diagnosis.
These results make a good case for using the technology as a tool to help doctors make decisions, but not as a replacement, as some might fear.
In a press release announcing the study's results, Dr. Marc Succi, MD, associate chair of innovation and commercialization and strategic innovation leader at Mass General Brigham and executive director of the MESH Incubator, said, "Our research thoroughly examined decision help using ChatGPT from the beginning of dealing with a patient through the whole care scenario, from differential diagnosis to testing, diagnosis, and management." "This shows that LLMs in general have the potential to be a useful addition to the practice of medicine and help clinicians make decisions with a high degree of accuracy," the statement added.?
The study, which was just published in the Journal of Medical Internet Research, is the latest step in the fast-moving relationship between healthcare and AI, and LLMs like the ChatGPT tool in particular. Some people worry that the technology could one day replace clinicians, but those who have been in the field for a while say it has value because it gives clinicians the knowledge they need to make decisions right at their fingertips. And the results of the study show in a subtle way that LLMs are good, but not good enough to replace anyone.
Officials at hospitals say they will do more research on AI tools like ChatGPT. For example, they will look into whether the technology can improve patient care and outcomes, especially in areas where knowledge and resources are hard to get to or scarce.
?