Though I don’t totally trust the generative AI wrt. generating arts (poetry, novels, illustrations etc.), the AI is somewhat useful like creating email to my supervisor or boss.
As my unreliable tutor, ChatGPT works well so far.
Because I have mental health issues, which may occur due to the heavy tasks like thesis studies, I need to ask for help from others to proceed with my studies. But there’s no one behaving like a personal tutor of linguistics in the rural city of Japan. Then, I decided to use Chatgpt in order to summarise my ideas and to get some tips regarding the study procedure (e.g. making a table of data, suggestions for the ToC or next steps I should take). This works well for me to get back on track with my thesis study, and now I have been able to finish each small task every day. Of course, I never searched for literature on AI or had them write a paper instead of me. (I made them write an email, tho. hehe)
The point is that I should not totally trust the data GPT gives me. While I have worked on linguistics, I realise their judgments on languages are not perfect like those of actual people. I sometimes say that generative AI is not always useful for language studies because of that.
In any case of using the AI, I would say that we have to trust our own judgment.
