Notifying users about the policy on AU concerning AI-generated content

Question

We have an opportunity to increase awareness of the Ask Ubuntu prohibition on AI-generated content. Currently we have only two active notification areas:

1. AI-generated content is not permitted on Ask Ubuntu: This is the policy itself which sits in Meta and will usually have a 'featured' tag affixed to bring a link to the main AU page.
2. Welcome to Ask Ubuntu: The 'Tour' page contains a link to the AI policy; in theory all users take the Tour but experience suggests that most do not.

It is a little scanty and there have been some users claiming ignorance of our AI prohibition when approached by Moderators about such content in their answers.

As a result of the consensus reached in the recent Moderation Strike there is another option that we can now bring in to play. This will entail a warning that will show to users pasting in content to the Answer box, notifying the user about the prohibition on AI-generated content. This could not be missed or overlooked.

On Stack Overflow this will look something like the following 'first concept' image:

My proposal is that we opt in and take this option with the obvious change of 'Stack Overflow' to 'Ask Ubuntu' and a link to our own policy, or to a new 'Help' page in the 'Learn More' link. But the beginning of this process, as with most larger changes, is consultation with the community, and here we are...

Answer 1

This question has now been available for comment for almost 3 months and has gained 2k views in that time. It has had the featured tag applied for most of this time to give greater prominence. The question itself has 23 positive votes and 1 negative as of January 26th 2024.

Discussion has taken place:

1. As comments and answers to this question
2. Informally on the chat room 'Raiders of the Lost Downboat' (chat room for AU editors and reviewers)
3. Informally on 'Ubuntu Moderation' (chat room for AU elected moderators)
4. Informally on 'AI Domination' (chat room for AI discussions)

Feedback has been overwhelmingly in favor of the addition of the following as a banner text option:

• **Reminder**: Answers generated by Artificial Intelligence tools are not allowed on [Site Name]. Learn more

Discussion has been a little divided as to the 'Learn more' link with some quite happy for a link to our own AI Policy on Meta while others have been happy with a 'Help Centre' page along the lines of the SO page.

I personally would be quite happy with an AU version of the SO AI Help Centre page that could be directly edited by the Ask Ubuntu Moderators. I have added a draft version of such an article here, it has drawn heavily from the SO help article and our own Meta post:

Draft version of AI 'Help' Page...

---

Why can't I use Artificial Intelligence tools to generate content?

Artificial Intelligence tools (a.k.a. ChatGPT, LLMs, generative AI (genAI) models, Google Bard, etc.) may not be used to generate content for Ask Ubuntu. The content (including questions and answers) you provide on Ask Ubuntu can only be:

• Your own original work
• The properly cited work of others (excludes citation of an AI source)

If the content you have placed on Ask Ubuntu is determined to have been written utilizing an Artificial Intelligence tool, that content will be deleted along with any reputation earned. Posting AI-generated content will lead to a warning from moderators, and any further infractions can lead to a suspension from the site.

What counts as “content generated by Artificial Intelligence tools”?

An “answer generated by Artificial Intelligence tools” is any answer crafted in part or in whole using a tool that writes a response automatically based on a prompt it is provided.

Because these tools are trained to answer with language that matches authentic text, the responses may look and sound plausible, but the quality of generated answers can vary significantly (up to, and including, completely wrong answers).

It is permissible to use grammar and spell-checking tools to correct grammar and spelling errors in your own posts or while editing and improving the posts of others.

Why am I not allowed to use Artificial Intelligence services to draft my answer?

Ask Ubuntu is a collaborative resource, developed and maintained by members of the Community. There are a few primary issues with content generated by large language models that make it unsuitable for use on Ask Ubuntu:

1. Users who ask questions on Ask Ubuntu expect to receive an answer authored and vetted by a human. This ensures that the answer is factual, relevant, and complete, up to the standards of another human. While human authors are not perfect, Artificial Intelligence tools may not take into account other important factors to a question (i.e., optimization, security, etc.), often add excessive noise to their answers (i.e., explaining all details, no matter how relevant), and may fabricate false or misleading information (hallucinations).

2. AI's cannot cite the sources of knowledge used up to the standards of Ask Ubuntu. Even when Artificial Intelligence appears to cite sources for responses, such sources may not be relevant to the original request or may not exist at all. For Ask Ubuntu, the answer may not honestly or fairly represent the sources of knowledge used, even if someone explicitly cites the Artificial Intelligence as an author in their answer.

What is an alternative to using artificial intelligence services?

Many of the answers on Ask Ubuntu are created by users sharing their own expertise, though this is not always required. In particular, users should cite trustworthy sources when supplying answers outside their expertise. Searching for sources, synthesizing them into a good answer, and citing them clearly are critical steps in developing a well-crafted answer.

---

Answer 2

There’s now an MSE post allowing us to opt in, which is probably the best way to do this unless we want something custom for some reason.

Sites can now request to enable a banner to warn about their policy on AI-generated content

Once we have consensus, a mod can give this question status-review.

Per this:

Both options have a "Learn more" link, which will point to a per-site help center article, whose contents should also be a part of the community discussion — this article should explain what the site's policy on AI-generated content is (here's SO's article, as an example).

It looks like we might need a help center page about our LLM-usage ban, too. IDK if we're allowed to link to Meta instead.

Answer 3

I agree with the direction, as I had proposed something in this regard in my post in January 2023:

Additionally, we could consider putting up small, unobtrusive banners where relevant (maybe accompanying the question- and answer submission forms).

While the message promises to be useful, I find a specific issue with the presented design mockup:

The problem is the placement of that banner, it's literally sitting inside the same textarea I am typing my answer into.

It seems to fall into a similar category with those distressing and "impatient" alert banners that some e-mail-address-field validation tools produce, while one is still typing:

"This is not a valid e-mail address! Fix it already!"

"I know, smart*ss, I have just started typing the first character, lemme finish first!"

You perhaps had already encountered such.

Now the hereby suggested placement of the banner has a somewhat related problem, and it is that this overly pesky banner™ keeps putting pressure on the user with a humiliating, and thus, highly distressing suspicion, that seems to presume at all times that the user wants to post AI content.

And this suspicion cannot even be resolved by correcting the content in any ways: I presume, the banner stays up all the while the answer is being produced.

One can get rid of that feeling of accusation only by completing the post and leaving the form.

That's humiliating and distressing.

Posting here ends up being rewarded with distress. We do not intend that.

---

It also does not make sense to put this accusatory piece of UI in between the content that I produce and the tools I use in producing it.

If I rely on any of those WYSIWYG buttons on the top, my mouse cursor has to travel over —and land right next to— this visually prominent, psychologically distressing ("no, I am still not committing that offense that it keeps reminding me about") piece of information.

---

I do believe that this effect is brought to existence solely by the placement, and proximity of that banner to the textarea device.

Putting the banner just a teeny-weeny bit away, could alter the message:

"We know that you wouldn't post that kind of content. This reminder has to be here out of necessity, but you are free to tune it out; your contribution is appreciated, and we are thankful for it; we recognize and cater to your need to post with dignity."

A different placement, just putting that banner a teeny-weeny bit away, could do away with the problematic effect.

Put it outside of the textarea-device please.

Answer 4

Thank you all for weighing in on this. The banner has been enabled, and the help center article has been updated.