Drupal Answers has recently seen a number of answers which are likely to have been created by generative AI (e.g. ChatGPT or another LLM). Content aside, the time span in which they were posted, and the length of the posts, leaves little doubt that they weren't written by a human.

The banning (or not) of such answers has been discussed at great length throughout the network over the last few months, the result of it being that each site needs to make their own decision about whether to allow generated content.

So: Should we allow content created by generative AI on Drupal Answers?

For reference, here is Stack Overflow's policy, which contains a lot of useful background, context, and debate about the subject. It's worth reading if you don't currently have an opinion on the matter (and probably even if you do).

Temporary policy: Generative AI (e.g., ChatGPT) is banned

There's also a list of how other sites are responding here:

Is there a list of ChatGPT or other AI-related discussions and policies for our sites?

For further reference, these are some of the generated answers. Please note There's no need to engage the OP of the posts any further about this on the main site, they'll be linked to this discussion and encouraged to participate.

  • We're allowed to define our own site-specific policy on this, we haven't done so yet most likely because this seems to be the first obvious pattern of it being used. Thanks for bringing it up. Are you ok me with editing your question a bit so we can use this as the post to create the policy? I'll just change a few words and add some links
    – Clive Mod
    Aug 30, 2023 at 19:47
  • Yes, that is fine with me. This is, by the way, the second batch of LLM posts I am aware of.
    – mona lisa
    Aug 30, 2023 at 19:48
  • That's good to know (well, good to know that it's been noticed anyway). Feel free to raise a flag on any other instances you see, the posts won't get any "extra" moderation until/unless this policy is decided, but it would be good to have them on our radar either way
    – Clive Mod
    Aug 30, 2023 at 20:24
  • There are some actions that are taken independently from how answers are generated and those actions may include deleting the answer. For example, if somebody answers more than a question with a copy-pasted answer, those answers are probably going to be deleted; if somebody keeps posting answers for the wrong question (answer for question A in question B, answer for question B in question C), those answers are probably going to be deleted.
    – apaderno Mod
    Aug 31, 2023 at 7:58
  • The same could happen for answers given without even understanding the question, which could also include answers posted after copy-pasting the question on an external site and using the output of that site as answer. This is not specific about LLM.
    – apaderno Mod
    Aug 31, 2023 at 8:00
  • Even in the case answers are not deleted, the community has still a way to avoid a user profile is used to continually post low-quality answers or answers given without even trying to understand the question.
    – apaderno Mod
    Aug 31, 2023 at 8:14
  • +1 for not allowing untested AI generated content.
    – Andreas
    Sep 18, 2023 at 15:00
  • 1
    There is now a settled policy upstream: Sites can now request to enable a banner to warn about their policy on AI-generated content - meta.stackexchange.com/questions/395881/…
    – mona lisa
    Feb 2 at 21:04

1 Answer 1


Please no. These answers are often just a list of possible solutions, untested possible solutions.

We have too broad question as close reason. And these answers are actually too broad answers and should be deleted because of this.

If you want that, ask ChatGPT. Here at Drupal Answers you get actual answers from real Drupal developers providing solutions that are tested and even get improved over time by them. They can relate to their answers because it's their answers.

  • I don't know what to think about generative AI-influenced answers. By that I mean, the community member did some research with AI, (importantly) tested the solution, then answered in their own words.
    – mona lisa
    Aug 31, 2023 at 16:21
  • @cilefen – That would be okay I think. They removed all the noise, did research themselves and provide a tested answer. That's +1 from me then. 😁
    – leymannx
    Aug 31, 2023 at 16:32
  • I really love this one: marketoonist.com/2023/03/ai-written-ai-read.html
    – leymannx
    Aug 31, 2023 at 16:35
  • 2
    This answer, and the first couple of comments, capture the whole thing perfectly in my opinion. If people can use LLMs for inspiration that's fantastic, it's just another tool in the belt. But these AIs are notorious for hallucinations, and are very good at forming eloquent, convincing prose that have no guaranteed basis in science. That's what we need to be careful of. If a human digests, verifies, and re-phrases a ChatGPT answer, it's pretty much the same as if they'd consulted a textbook
    – Clive Mod
    Aug 31, 2023 at 20:18

You must log in to answer this question.

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