4

I recently stumbled with this answer and some aspects of it raised my suspicion that it is AI generated, albeit with slight modifications made by the poster. The answer in itself is highly technical and detailed, formatted with AI bullet points style.

And then a user comments highlighting factual errors of what is supposedly be knowledge from an expert.

Browsing the user's who posted the answer StackExchange profiles, nothing points to such expertise.

Meta discussion about the subject, accepted and most voted answer:

In short, I think we should handle ChatGPT answers the same as we would handle copy-paste from other websites without attribution, but with added prejudice because of: wasting everybody's time, the difficulty of detection, and the need to dissuade other people from the easy "rep-farming" that will occur if we hold a tolerant stance on ChatGPT answers.

Should suspected AI generated answers be flagged for moderators attention?

1
  • Naughty Schrôdinger ! Leaving to mods the responsibility for opening the box… ;-)
    – MC68020
    Commented May 10 at 23:29

1 Answer 1

5

Yes, please flag answers that you suspect do not follow the current policies of the site.

As for the specific answer that you are highlighting, I'm not certain enough that it is AI-generated to take action against it. It is problematic that the author of the answer has been made aware of issues in the text without updating it, but this alone does not mean that it is an AI-generated text.

Using an enumerated list for the structure of the majority of the text does make warning bells sound, but this is also how some users prefer to present their text.

The style of writing, in terms of expression (it's personal, detailed, and direct), use of markup, and the fact that there's a list of external references at the end, makes me think it's not a generated answer (even though more recent ChatGPT models are able provide links to further resources). Also, when prompting an LLM about the issue at hand, I could not generate any text that looked similar to the posted answer.

AI-generated answers seldom see updates from their users, and when challenged about the origin of the text or its contents, they seldom reply at all. When made aware of the site's policy on AI-generated posts, some users promptly delete their generated answers, and others let us do that work without grumbling too much about it.

You must log in to answer this question.

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