I have read Stack Overflow's current stance on AI generated content and understand the decision making process that went into deciding that AI generated content was not suitable for answering questions.

The post clearly states "All use of generative AI ... is banned when posting content on Stack Overflow", yet I can't help but wonder if this should apply to writing questions. I've run into a lot of poorly written (bad spelling, grammar, poorly formatted code...) questions, which I totally understand. Not everyone is writing in their native language or is familiar with code formatting and can't be expected to get it right (at least when new to the site), but I'm wondering if this could be an appropriate case for AI, to basically improve the quality of questions without impacting the actual question being asked. My questions then are:

  • Does the Stack Exchange have a set policy on using AI for writing questions?
  • If not should I encourage someone to use AI when writing their questions?
  • Are you asking about the Stack Overflow site or all sites on the Stack Exchange Network?
    – Makyen
    Commented Feb 27 at 6:18
  • It was a general question although the information I had on the issue was more specific to SO. Also is there a particular reason for the downvotes?
    – 35308
    Commented Feb 27 at 6:25
  • 6
    I didn't vote, but I'd suspect that the downvotes are because the issue you're asking about has been discussed many times over the last year, including both all of the specific points you've brought up and the spectacular failure of SE's own attempt to use generative AI to "help" questions. The downvoters probably believe that you asking this question indicates that you didn't do much research about this issue prior to asking this question. Admittedly, it's possible that you did do research and just didn't find anything. OTOH, people are tired of rehashing the same points.
    – Makyen
    Commented Feb 27 at 6:44
  • 4
    I also didn't vote, but adding to @Makyen - Downvotes on Meta (i.e., here) can indicate disagreement, so disagreement with encouraging AI usage is a possibility. Also, while it isn't your fault, the Stack Exchange community has strong opinions and a messy history with generative AI on the platform. In brief, there were a number of poorly-received genAI initiatives by Stack Exchange and a moderation strike heavily involving genAI policy.
    – cocomac
    Commented Feb 27 at 6:46
  • Sadly, you won't get an official answer. It is simply too easy to pretend that the communities will self moderate and have them get the blame for any decision. You may look at a volunteer moderator answer here for reference but that is probably all you will get as "official" goes. Also, have an upvote even if this was mentioned before, exactly for the reason I just mentioned. Commented Feb 27 at 10:03

2 Answers 2


Tl;dr: do not do that. Policy aside, genAI to "improve" questions is going to go poorly, as demonstrated here.

I'll do a frame challenge, and explain the problems with using genAI here.

Stack Exchange, Inc. (the company behind Stack Overflow) tried to introduce a formatting assistant powered by genAI, and you can see how that went. See also this Meta Stack Overflow post. If I understand what you're asking, it's similar to what they did, but just manual.

Mithical's answer includes the prompt they used, which was intended to improve Stack Overflow questions. Unfortunately for Stack Exchange, trying to ask an LLM has a number of issues, and you're likely to face those same issues too:

Most of those problems (except for supporting the mod strike!) are challenges you're likely to encounter if you use ChatGPT or similar to "improve" your question. Other options include:

  • Spell checking tools
  • Relying on community editors to fix minor grammar errors in your post

Stepping away from the frame challenge to actually answer the policy question, it's somewhat banned. The MSO ban would likely ban it in its entirety, but on sites without an AI policy banning all genAI content, it would only be banned if uncited/unattributed, per the Inauthentic Usage policy guidance.


Does the Stack Exchange have a set policy on using AI for writing questions?

Stack Exchange does not have any network-wide ban on using AI for writing questions. The policy is decided for each Stack Exchange independently.

If not should I encourage someone to use AI when writing their questions?

Assuming the use of AI is allowed, that depends on whether it helps them write better questions.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .