I brought this up here. Wondering if any users have posted obvious AI-generated text on articles and what, if any, policy on AI usage MA ought to have. (I expect it would be disallowed.) This will become a problem eventually. Which "Memory Alpha:____" page should state the policy, or does it deserve its own page? --LauraCC (talk) 03:40, 25 April 2024 (UTC)
- I'm not aware of any AI-generated text on articles, but it could have slipped through the cracks at some point.
- I think MA should have a policy: no AI-generated text. The main issue with AI text is a legal one: if the text is truly "original" and not derivative from copyrighted text, it would be in the public domain (like all material not made by Humans). Therefore, we would have to mark the text as public domain (which would be cumbersome and pointlessly complicated) so it is not assumed to be licensed under the CC BY-NC. If, however, it is derivative, we would be using text without permission, which would either qualify as fair use or a copyright violation. If the use qualified as fair, we could get away with it, so long as we, again, marked the text as such and identify the owner. If the use didn't qualify as fair, we would obviously have to remove the text. None of these cases are ideal, and one is illegal. Since it can be difficult to know for sure which of these cases is the one you have with a given portion of AI text, the easiest and most logical solution would be to ban all AI-generated text.
- Now, the next question is where to put this policy. For that, we have two options: (1) add a clause to an already-existing policy page, such as MA:CYS; (2) create a new policy page. I recommend the second option.
- If we do indeed create a new policy page governing AI-text, we should name it something like "Memory Alpha:Artificial intelligence" or "Memory Alpha:Using artificial intelligence".
- This doesn't seem to be an immediate concern, as, to my knowledge, MA has never had problems with users adding AI-generated text. However, I agree with LauraCC: this is still an important topic that we should address sooner rather than later, as it is bound to become an issue eventually and it would be a good idea to "lay down the law" so less people, if any, choose to add AI text. I'll start writing a draft policy as soon as I can, and if people approve it, we can turn it into an official policy.
- Sorry for the wall of text ;) 🖖 Mr. Starfleet Command (talk • contribs) 17:11, 25 April 2024 (UTC)
I think it should be its own page too. Otherwise, it would be easier to ignore.
In addition, also add language about AI generated images. Also, provide tips for recognizing AI generated text (I've heard some words are more common to AI text than general usage, supposedly, with "delve" being one example.) And of course, AI generated images have been noted to be uncanny, e.g. having grotesquely unrealistic hands. What's a reliable source for these guidelines? --LauraCC (talk) 17:18, 25 April 2024 (UTC)
Also, add "AI generated" to the Memory Alpha:Deletion policy page list of reasons to delete whole pages and/or images. --LauraCC (talk) 17:40, 25 April 2024 (UTC)
- While we may not have a long-standing policy, it's on Fandom's radar and I think they're developing guidance, landscaping how wikis may be using it now (and legal implications.) I think the best course, rather than develop a new policy page which risks being slow to adapt to changes in Fandom policy/industry capability/law, would be to include a broad snippet in existing content policy. Something like, "Use of AI in writing article content is discouraged and may violate Memory Alpha's and/or Fandom's licensing and copyright." Maybe "MA is a community, intended to be written collaboratively by its readers," or something less flowery.
- Does our File use policy not already essentially rule out AI-generated art? I suppose we could add a line there to be more direct, but most acceptable files as is are screenshots or published media cover-type things. - AJHalliwell (talk) 21:21, 22 June 2024 (UTC)
- In response to Laura's suggestion above, Sulfur added "Images that are clearly and obviously AI-generated" to the list of immediate delete reasons, so that part's taken care of already.
- I agree with your suggestion (especially because I haven't even started a draft policy yet ;) ), although I'm not sure where such a note would best be placed. Perhaps on the copyrights page? 🖖 Mr. Starfleet Command (talk • contribs) 21:31, 22 June 2024 (UTC)
"As Star Trek fans, we are all well aware that any new technological advances have the potential to malfunction or be misused/abused. With that in mind, we aim to proceed with caution and due consideration before making use of AI technology on Memory Alpha. Until such time as policies and procedures have been put into place for acceptable usage of these programs, please refrain from adding AI content to this website."
Friendly, and cites a common Trek trope as justification (tech gone wrong). Thoughts? -LauraCC (talk) 18:12, 26 June 2024 (UTC)
Not an article, but the subject of AI just came up: Talk:Pathway drive -LauraCC (talk) 00:04, 18 April 2025 (UTC)
I think this page may be AI generated too: Space travel effects. -LauraCC (talk) 16:08, 13 September 2025 (UTC)