7

I've heard that some sites, including Stack Overflow, have banned AI-generated answers.

Are AI-generated answers, such as those created using ChatGPT, permitted on this site?

1

2 Answers 2

Reset to default
9

Tl;dr No. AI-generated answers are prohibited on GenAI Stack Exchange. Note that this also applies to Tag Wikis. AI-generated content is still prohibited even if is listed as AI-generated.

Answers created by LLMs are generally harmful to the site. They carry a number of significant issues, but the "primary problem is that while the answers which ChatGPT and other generative AI technologies produce have a high rate of being incorrect, they typically look like the answers might be good and the answers are very easy to produce." [1]

The ease of generating that type of "answer" often results in users posting a large volume of AI-generated answers, and they usually require an SME to determine if the answer is actually wrong.

Note that this doesn't mean you may not include AI-generated content as examples to support your answer - it means that your answer itself must not be AI-generated.

[1] Temporary policy: Generative AI (e.g., ChatGPT) is banned by Makyen◆

[2] Announcement: AI-generated content is now permanently banned on Ask Ubuntu by andrew.46◆

8
  • I'm interpreting this answer as being only applicable to wholly AI-generated answers (e.g. blind copy/pasting from ChatGPT), not to AI-human coauthored answers.
    – Rebecca J. Stones
    Commented Jul 18, 2023 at 13:49
  • 1
    @RebeccaJ.Stones What do you mean by “AI-human coauthored answers”?
    – wizzwizz4
    Commented Jul 18, 2023 at 17:35
  • For example, someone writes an answer and asks GenAI to proofread it, then modifies their answer based on its feedback (maybe the author missed something important). Or perhaps they use GenAI to save time generating a plot for their answer. Or uses GenAI for grammar fixes or translation. Or they got help with their initial brainstorming.
    – Rebecca J. Stones
    Commented Jul 18, 2023 at 22:00
  • @RebeccaJ.Stones We should understand the prohibition set in this answer is referring to a "Generative AI answer" posted as is, without any human verification and validation. The spirit of this prohibition is that in any way, a human being might excuse themself for post inaccuracies and consequences. Any post responsibility and liability belongs to the flesh and blood person behind the post. In other words, if the post is unhelpful, or wrong, and this leads to a post ban, then the human behind the post should consider themselves post-banned.
    – Wicket
    Commented Jul 19, 2023 at 5:10
  • I'm not saying we are a "criminal court" or something like that. Still, posts that look like an AI hallucination should be deleted immediately (actually, as soon as the moderation strike ends, giving a proper time to handle the backlog).
    – Wicket
    Commented Jul 19, 2023 at 5:19
  • 3
    @Wicket It's not just about responsibility. GPT systems write the kind of text that my eyes slide off; unless I'm completely alert, I'll often miss very incorrect things. By the time we discover the wrong information that's written, it's already too late: people will have read it, and repeated it. This is tolerable on a small scale, but not on a Generative AI scale.
    – wizzwizz4
    Commented Jul 19, 2023 at 7:53
  • @wizzwizz4 I agree with you. We need to emphasize the magnitude of this problem loudly and use any available platform to draw attention to it. We need to make it more visible that the people participating in moderation and content curation are insufficient despite many participating in the moderation strike.
    – Wicket
    Commented Jul 19, 2023 at 8:02
  • This was posted today by Philippe, VP of Community for Stack Exchange Network (Interim) Policy on AI-content detection reports.
    – Wicket
    Commented Aug 6, 2023 at 21:47
-3

It is acceptable to post AI-generated answers. We shouldn't blindly discriminate against AI. If some AI provides a correct answer, possibly with a few modifications made by a human, then it shouldn't be deleted on the grounds that it was AI-generated. No need to waste human time if some AI can do it. However, if an SE user tends to post incorrect answers, then they should be banned.

1
  • 2
    I understand Stack Exchange requires GenAI usage to be acknowledged, although I think this rule applies to e.g. ChatGPT-generated content, rather than autocomplete, spelling and grammar checkers, speech-to-text, Google Translate (?), and other such AI-generated content.
    – Rebecca J. Stones
    Commented Jul 23, 2023 at 23:04

You must log in to answer this question.

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