Table 2.

Clinical use cases endorsed by survey respondents who indicated that they had used ChatGPT or a large language model (LLM) in their clinical work (n=52).

Responses to “How have you used ChatGPT/LLM in your clinical work (select all that apply)?”	Respondents, n (%)
Draft school or work letter	26 (50)
Draft prior authorization	18 (35)
Generate patient education materials	15 (29)
Generate differential diagnosis	12 (23)
Other	12 (23)
Ask a specific clinical question (not mentioned above)	11 (21)
Draft all or part of a clinical note	8 (15)
Suggest a treatment plan	8 (15)
Respond to patient inbox messages	5 (10)
Draft all or part of a discharge summary	3 (6)
Draft handoff documentation	3 (6)