|
How ChatGPT Became a Standard to Evaluate LLMs in Healthcare
(Thursday, May 16, 2024) Over the last year, numerous researchers have evaluated the potential of ChatGPT in healthcare applications from writing medical notes, to diagnostic decisions, prescription recommendations, and even medical consultations and research. However, ChatGPT is but one of the various large language models (LLMs) being evaluated in healthcare. And they all are in their infancy. ChatGPT has become the representative of all artificial intelligence (AI) based algorithms for doing semi-intellectual tasks. It is unquestionable that ChatGPT can expedite and standardize several routine tasks. It can help write better medical informed consents, help author manuscripts, educate patients, or strategize drug development plans. The word “ChatGPT” is used the same way we use “Xerox” when talking about photocopying. But that’s where the challenges with over-relying on ChatGPT exist. When conducting an experiment to evaluate the potential of LLMs using ChatGPT, it is easy to generalize the lessons to all such tools. Often the reports from such research make broad claims about the limitations, benefits, and risks of LLMs in general based on an experiment where ChatGPT was the only tool used. This limits the value of such research. New tools are being developed to address the limitations of ChatGPT. And even ChatGPT is evolving. The initial research is just an indicator of what is to come in the near and far future and should be treated as conclusive. In our attempt to hype the potential of LLMs using ChatGPT, we may unintentionally hurt the potential uses of LLMs overall. A note of caution. 
AUTHOR
Dr. Mukesh Kumar Founder & CEO, FDAMap Email: [email protected] Linkedin: Mukesh Kumar, PhD, RAC Instagram: mukeshkumarrac Twitter: @FDA_MAP Youtube: MukeshKumarFDAMap |
|
|
|
