SAN FRANCISCO—A study tasking artificial intelligence with deciding whether or not performing an epidural is safe, based on when patients last received anticoagulation, found that large learning models (LLMs) are unable to accurately provide clinical decision making.
“LLMs such as ChatGPT or Bard have recently emerged as a technology that can produce very authentic-sounding discussion,” said Nathan Hurley, MD, PhD, an anesthesia resident at the University of Wisconsin School of