Wiki and Forum Policy - General Discussion

IIRC, its because we have the Image Quality Upgrade thread for fixes to otherwise good pics and image quality alone isn't enough to pull an image unless it makes it hard to see the trope clearly or exceedingly bad. In that case, the thread can easily be folded under "Needs Help" or "Does Not Illustrate" (if the quality is so bad the trope can't be seen clearly).

Yep, all of that.

Edited by Willbyr on May 8th 2024 at 8:49:28 AM

Remind me, as far as website is concerned, copying text from sources that explicitly permit sharing even with credit still counts as plagiarism? Recent ATT made me question if websites having CC license has any effect on copy-pasting, which I assume would still not be allowed.

Other sites' policies don't matter to us in this regard. No matter what license is present on content, we must follow the rules of Fair Use, specifically excerpts and attribution.

Any content reproduced in whole must be indicated as such, may only be done with the permission of its author (and concurrence of the staff), may not be altered, and must be attributed.

This isn't about copyright per se; it's about plagiarism. We can't pass off other people's work as our own, under any circumstances, for any reason.

Edited by Fighteer on May 10th 2024 at 12:49:57 PM

one question, what happens if I report a page by mistake as in I misclicked the "Report a Page" button.

I can't speak for the whole team, but if I were checking on hollers and saw a holler that was sent by accident (the Report Page button's reports are identical to hollers), I'd just ignore it, so nothing would happen.

Edited by GastonRabbit on May 16th 2024 at 3:14:34 AM

As this has been brought up, again, a question.

There was a discussion here if AI can be used for troping, which stalled into "as long as edits are policy-compliant, but we don't believe AI can reliably understand our policies or accurately describe works without making stuff up". This was also up to consideration to append into Please Trope Responsibly along with automation scripts and grammar plugins.

Since AI are getting better and with allowance of AI images this topic also got brought up, does the site have a stance on AI-tools for wiki editing, or it's "at own risk"? The question at hand is about using AI to generate summaries for works, but this can easily carry over into writing examples.

Edited by Amonimus on May 18th 2024 at 4:16:29 PM

I've seen people admit to using AI to help fix the grammar on their own examples. It's not like we can really police it easily, so I would just judge it like we would any human-written example, with those using AI accepting the usual risks that it's a machine unfamiliar with troper conventions beyond what it can glean from a bunch of scraped contextless data.

Speaking from the heart, with only a smidge of sarcasm, it would be hard for AI to do worse at editing this wiki than some tropers I've seen.

Judge the content, not its provenance.

Ouch, man.

I think it would be better to teach tropers to use AI as a tool, and use it responsibly. Don't just generate and post, check the text for factual errors, rambling, that sort of thing.

AI can be a very powerful tool for a wiki, but users have to learn how to use it, and how not to.

For instance, would AI be useful for crosswicking? That only involves rewriting bits of text into grammatically correct paragraphs, after all, which AI is already pretty good at.

Yes, it doesn't really matter, as long as the troper does it responsibly, in which case you couldn't really tell anyway.

But I think we could use a policy page on how to use AI responsibly for editing.

Edited by Redmess on May 19th 2024 at 12:02:28 PM

I'd rather not have a repeat of an earlier debate and focus on the wiki. Am I to understand that as long as examples are accurate it doesn't really matter "who" writes them, and if they are in fact inaccurate they will be treated similarly to when malfunctioning nunny plugins warrant an immediate suspension (for reference, that part is included in What to Do If You Are Suspended)?

Edited by Amonimus on May 19th 2024 at 12:59:42 PM

I think it's fair to say that if you delegate formatting, fact-checking or any other aspect of editing to an automated tool, you're responsible for ensuring that it doesn't harm the wiki.

And if it does there may be consequences, as suggested.

Again though I'm not too sure what there's to do here; I'm fairly certain Kory's rule covered all AI content.

Well, we just have to make sure it is clearly laid out in the rules, and a support page explaining how to properly use AI would be handy to point people to who find themselves in the EB thread over AI issues.

You can't really make a guide on that? The point is that people should double check what they add on the wiki AI or not beforehand regardless of what tools they use.

Good edits are good edits no matter how they are written, and tropers are responsible for their edits, ditto. We need to stop freaking out about this.

I'd actually be somewhat curious to see whether a generative AI could write a trope article for a work that's sensible. I highly doubt it.

As long as the work is known to the AI, it may actually do a pretty good job if you know what questions to ask (provided you know the work so you can check, of course).

Unfortunately, there doesn't seem to be a clear way to check whether or not an AI has a work in its database, at least not with Chat GPT.

Maybe one day we can make a TV Tropes dedicated AI trained on the wiki and forum, where we can upload new works into the model as needed.

Edited by Redmess on May 19th 2024 at 5:59:16 PM

If you use it to generate text, then yes. Maybe. But any time you are using it to get information, it will immediately fall apart, because AI has no ability to put any value judgment on what information is right or wrong. It will just push out a bunch of gibberish about how R2-D2 is a unique model because all those words are in its database and it has no way of telling why this might not be correct.

AI is designed to generate text that could conceivably have been written by a human. That's it. If that's all you want it for, fine, it works. But trying to use it to seriously answer questions will immediately run into factual flaws.

Which is all fine and dandy for our purposes, because it doesn't matter who or what wrote the examples, we can judge them on their own merits. But it's a problem for the people trying to use the AI to write examples. If they keep adding incorrect information that other tropers have to clean up, they'll eventually be banned, no matter where that information came from.

Fighteer posted in the Legal Eagle thread that:

and I’d like to confirm what level of semi-public content has to be to qualify under this rule? Obviously TV shows and Films don’t count (though with staggered film release schedules I could see room for halting discussion until global release) and based on Fighteer’s ruling You Tube creators who post their content elswhere does count, but what about related content that is exclusive? The Real Life Lore You Tube channel has a series of Nebula exclusives, are they not allowed to be discussed on TV Tropes? Jet Lag the Game does an exclusive podcast alongside the main episodes Nebula release, with the main episode getting a You Tube release a week later. I’m assuming the intent isn’t that we can discuss the podcast but not the main episode but I think clarity is needed.

Likewise how does it work with Livestream content, do we have to link to the livestream with a timestamp? What if the Vo D is restricted to paying subscribers? What if the Vo D is sometimes put on You Tube later but not always?

Finally, are there other platforms covered by this rule than just Nebula and Patreon? I assume that You Tube early releases are also covered?

The new rule seems to have created a lot of questions so some clarity would be helpful.

Edited by Silasw on May 24th 2024 at 7:54:36 PM

This is the first time I hear about it. We've discussed several times that any form of non-restricted release is as good as a full release (paywalled content like Patreon also deemed effectively public) and is tropable, so it not being allowed to be discussed would be odd, though makes some sense. I could argue such content should be heavily spoiled or avoided in forums unless the thread explicitly allows otherwise, though, as there's a divide between people who are waiting for it to get more public.

e:

Oh, ok.

Edited by Amonimus on May 24th 2024 at 10:31:19 PM

To me it looks like Fighteer was singling out forum discussions of paywalled content, not whether it can go in articles

We had this discussion a few times in the Media forums, when something gets early access release on Patreon and later full release on YouTube or whatever. It is discourteous to discuss such content before its public release.

This is what we say on Creating A Work Page For An Unreleased Work:

That's guidance for wiki rather than forums, but I think the general principles are going to be consistent. We are a little more flexible on forums — for example, we'll allow content leaks to be discussed behind spoiler tagging — but this is our starting point.

Edited by Mrph1 on May 24th 2024 at 8:27:19 PM

Not a mod, just clearing up a little confusion that's come up in the past regarding the paid subscription part.

IIRC, the reason for this is that it's too hard to verify information from subscribers-only works. The work that led to this policy was RWBY when it had a one-week subscriber early access, leading to questionable examples being added every week and "normal" tropers not being able to access the work to verify them.

Ah, good. Glad to see I made the right call when I decided to avoid troping the Patreon-only episodes of Ain't Slayed Nobody.

I'm not sure I understand this rule. We have no such restrictions when a show runs on a streaming service, which is just as exclusive as Nebula or Patreon. This distinction seems somewhat arbitrary.