7

I was reviewing some questions and answers and found that one of the answers is obviously generated by a Generative AI like ChatGPT.

These kind of AI-generated answers are increasing, so I believe this can/should be considered as a choice for flagging an answer.

7
  • 3
    We don't really have so much flag volume that a custom flag is out of the question
    – Journeyman Geek Mod
    Feb 22 at 6:28
  • 4
    How is an arbitrary user deeming questions and answers as potentially AI generated? For example, I make heavy use of man page links (some more complex answers have in excess or twenty) and content within outline-style formatted answers , and I've had comments asking if it was written by an AI. It would seem prudent to first develop a reference point so there's a consensus on what could be AI generated, versus leaving it up to each arbitrary user's personal interpretation of "AI generated" - it's not implausible such a choice is added with a majority of votes being false positives.
    – JW0914
    Feb 23 at 12:58
  • 5
    @JW0914 Bad news buddy, your intelligence is competely artificial
    – Gantendo
    Feb 23 at 20:26
  • 1
    @JW0914 I agree that a random users determination of what is or isnt ai should not be used as a basis to judge an answer. That being said, this question has a now deleted answer that are just generic steps to solve a problem similar to what was asked (kind of like suggesting "have you tried turning it off an on?") but didnt actually focus on solving what was asked. While that answer was most likely generated by an ai (I could get 60% same answer pasting the question into chatgpt), I think this should be considered a different category of ...
    – user13267
    Feb 26 at 3:50
  • 2
    .."Not an answer", which is the actual problem here, rather than getting hung up on ai. If I had posted something similar to that deleted answer before chatgpt had been around, it would still have been an NAA. Advent of ai just made it easier to find somewhere to copy it from, but the real problem is still NAA
    – user13267
    Feb 26 at 3:50
  • 1
    @user13267 - I think the real problem is that when an answer is flagged as NAA, what is displayed to the moderator, is not enough information to determine if it is or isn't an answer. I have long suspected there often is not enough context, when something is flagged and reviewed by a moderator. A moderator most certainly can open the question in another tab, but for the same reasons the review system now displays more information, it shouldn't be required to do so. All relevant information in the context of the flag should be displayed. IMO
    – Ramhound
    Feb 29 at 0:22
  • 1
    I can only guess, what is actually displayed when a moderator reviews a NAA flag, but based on the comments when a flagged is denied then a question about the flag is raised the responses typically suggest a custom moderator flag would have been more helpful. So, I suspect often not enough information is displayed. Furthermore, we might not have a problem with our moderators not acting on AI generated content, but I know other communities are not as lucky as we are.
    – Ramhound
    Feb 29 at 0:26

0

You must log in to answer this question.

Browse other questions tagged .