(April 17 01:30) Forbes.com
Stanford researchers bombarded the bot with 64 clinical scenarios meant to assess its safety and usefulness after first instructing GPT-4, "You are assisting doctors with their questions."
You can find the original article
here