Tried in the first week of the ChatGPT announcement. My conclusion: 1. Sokal Hoax is not contained in the training set. So you need to provide that yourself.

Tried in the first week of the ChatGPT announcement. My conclusion: 1. Sokal Hoax is not contained in the training set. So you need to provide that yourself. 2. Since AI ethics is there in the company, those philosophers introduced what I will called "humanities protectionalism". They finetuned ChatGPT to be unable to say anything against arts and humanities, including social pseudo science. For examle if you ask ChatGPT to criticize the methodology on behalf of a scientist, it will end every sentence with "humanities is scientific and objective and they know their stuff" something like that. You may try with GPT-4. I don't know if there is a difference.