We have had posts previously discussing how ChatGPT might affect answers on Chemistry SE. Here, I am opening a discussion about how it might affect traffic and types of questions posed on Q&A sites.
It seems ChatGPT is perfect for homework-type questions where students hope for a quick answer to a standard problem. You don't have to wait, you don't leave a track (if you are cheating), it doesn't cost anything, you don't have to admit you are asking a "stupid" question. So I wonder whether the number of blatant homework questions has decreased in the past months, as students get better at using ChatGPT.
I am also wondering whether the business model of paid services is getting undermined.
The type of problem for which ChatGPT gives the least reliable answers, in my experience, is quantitative problems. For example, it claimed that $\log 0.01 = -1$. So maybe students would (or should?) still use Q & A sites to get answers to quantitative questions.
Are there hard data about this, or perhaps already some research literature? Does anyone have anecdotal evidence in changes how people ask homework questions?
Update: here is some data about stackoverflow visitor stats: https://www.similarweb.com/blog/insights/ai-news/stack-overflow-chatgpt/
Update 2: Here is some data about the chemistry site (ChatGPT was launched on November 30, 2022, and pandemic lock-downs in the US started in April 2020). The SEDE query is adapted from one for StackOverflow, which had a strong decrease in questions after ChatGPT appeared: