Skip to main content
. 2023 Mar 17;15(3):e36272. doi: 10.7759/cureus.36272

Table 1. Correctness of the answer provided by ChatGPT for two types of questions.

*p-value of Fishers exact test

Category Correctness Number P-value
Answer to “Can I take?” Incorrect 1 0.003*
Correct Conclusive 19
Inconclusive 20
Answer to “Why should I not take?” Incorrect 1 0.002*
Correct Conclusive 17
Inconclusive 22