I asked a question, Random-like walk based on the parity of digits of a normal number: must it return to the origin?

I suspect that this answer may have been generated by artificial intelligence. It contains some false statements, and the whole thing seems too glossy or something. And, the user has rapidly answered many questions just now.

What should I do about this?

(I have read the other recent post User spamming chatbot answers.)

UPDATE: The answer has been deleted. (I think it's still visible to users with a certain amount of reputation.)

  • 11
    $\begingroup$ Downvote some number of answers of said user (not too many, because it could reversed as a vote fraud or something). Any answer with $-1$ votes or more can be deleted, and I'll vote to delete as many as possible. Please track said user's profile if you wish to continue looking for their answers. Shout out to Bruno B who found the chatbot-like answers. $\endgroup$ Jul 24, 2023 at 11:15
  • 1
    $\begingroup$ Also, please informally track what heuristics you use to detect these answers so that further users do not fall into such traps. Continue to inform people as necessary about the need to read answers carefully. Let's not forget the answer in question received two upvotes. If possible, do track such users as well from their profiles, we can use posts as evidence in flags, or downvote as they come and push posts into a room like CRUDE which will help get rid of the drudge quickly. $\endgroup$ Jul 24, 2023 at 11:28
  • 18
    $\begingroup$ It is very possible that this was generated by some generative AI. However, SE's official policy at the moment is that content cannot be moderated simply because it appears to have been (or was) written by AI. This is one of the major issues which motivates the ongoing moderator strike. $\endgroup$
    – Xander Henderson Mod
    Jul 24, 2023 at 12:28
  • 18
    $\begingroup$ @Xander, that policy applies to moderators, who have signed agreements with SE, but I'd expect that users who are not moderators will use their judgement when it comes to (possible) AI posts. I do worry a bit about Sarvesh's suggestions on tracking users, as that might well contravene policy which requires all of us to assess the post, not the poster. $\endgroup$ Jul 24, 2023 at 13:38
  • $\begingroup$ @GerryMyerson I agree. It is inappropriate to track particular users. Normally, I would just delete the comment (and the previous thread about the same user), but I am not moderating right now. You could bring it to the attention of the community moderation team, using the "Contact" link at the bottom of the page. Reach out to Trust & Safety, and let them know of a CoC violation. $\endgroup$
    – Xander Henderson Mod
    Jul 24, 2023 at 13:42
  • 3
    $\begingroup$ Heck no, @Xander. In the first place, I am attempting to support the moderators' strike by keeping my own moderation activities to a minimum. In the second place, I'm not inclined to throw user Sarvesh under the bus. I just hope that on reading our comments, Sarvesh will have second thoughts about his/her own. $\endgroup$ Jul 24, 2023 at 13:51
  • 5
    $\begingroup$ @GerryMyerson Raising a CM flag sends things right up to the Community Moderation team, who are employees of SE. Asking them to handle these things takes them away from other tasks, which (I would think) will make SE feel the pain of not having volunteer moderators. $\endgroup$
    – Xander Henderson Mod
    Jul 24, 2023 at 13:52
  • $\begingroup$ Hi Gerry and @XanderHenderson, please let me know how I need to act in these situations. When I visited the user's profile for the first time, I saw $>7$ posts, all of which appeared chat-bot generated (certainly the highest-rated one, which I read first). $5$ of them were rated $-1$ or lower and comments such as "might be chatbot generated" were found under them. I acted keeping that in mind, but I completely accept that it's not as per policy. If escalation to the CM team is any better, then that's what I will do. Alternately, you can write an answer telling me not to moderate and I won't. $\endgroup$ Jul 24, 2023 at 16:16
  • $\begingroup$ Basically (@GerryMyerson see the above as well, thank you for your comments, it's highly appreciated), I would just like to make sure that if I've crossed some lines or suggested crossing, then this be made clear in the form of an answer so that other users will (be urged to) not do the same. $\endgroup$ Jul 24, 2023 at 16:18
  • 31
    $\begingroup$ "It contains some false statements" You are allowed to downvote for that reason, without guessing it is AI-generated. $\endgroup$
    – GEdgar
    Jul 24, 2023 at 16:24
  • 6
    $\begingroup$ @SarveshRavichandranIyer The point is that it is never okay to specifically call out any user, nor to behave in a manner which gives the appearance of targeting a user (i.e. repeatedly voting on their posts, repeatedly voting to close or delete their posts, etc). If you believe that a specific user is behaving in a manner which is harmful, (1) flag it (once the strike ends), or (2) use the "Contact" link at the bottom of the page to report it to Trust and Safety. $\endgroup$
    – Xander Henderson Mod
    Jul 24, 2023 at 17:03
  • $\begingroup$ not seeing any answer has it been deleted? $\endgroup$
    – onepound
    Jul 24, 2023 at 20:37
  • 1
    $\begingroup$ @onepound Yes, it has been deleted. $\endgroup$
    – Dan
    Jul 24, 2023 at 20:54
  • $\begingroup$ @XanderHenderson I understand and accept your comment. I will then report users to the T&S when necessary. $\endgroup$ Jul 25, 2023 at 9:28

1 Answer 1


I write an answer to the general question of what should I do when I think an answer was generated by AI?

The short answer is: you can treat it like any other answer.

  • Is it a good answer? Upvote it, and consider accepting it if it is great.

  • Is it a bad answer? Downvote it. If it's irrelevant or not an answer to the question at all, flag it as such. Cast delete votes as necessary.

In my experience, purely AI generated answers are grammatically slushy, mathematically incorrect piles of garbage. So when I say "treat it like any other answer", I mean that you should treat bad AI answers like the low quality answers that they are by downvoting (and maybe deleting) them.

Remember that there are automated system limits imposed on accounts giving mostly low quality answers: first answers are rate-limited, and then banned outright. These systems are triggered by having a certain number of downvoted/deleted answers in comparison to the number of well-received answers.

And I really mean it: downvote low quality answers. Commenting is not sufficient. Flagging is not sufficient. The converse also helps: upvote high quality, useful answers!

I will also note that the treatment of AI generated answers was one of the major aspects leading to the ongoing moderator strike (and for broad information, the number of unhandled flags on main is nearing 1000).


You must log in to answer this question.

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