[WolfCop]

From the paper:

> ...the original full text of the question was put into a fresh chatbot session, in which the session was free of prior questions asked that could bias the results (version GPT-3.5, OpenAI), and the chatbot response was saved.

It seems like they just pasted the question in. For those who have asked it for medical advice, how did you frame your questions? Is there a prompt that will help ChatGPT get into a mode where it knows it is to provide medical advice? As an example, should it be prompted to ask follow up questions if it is uncertain?