Topic: Request to create a "Is this AI?" thread to avoid unnecessary use of the flag tool

Lately, I've noticed an increase in the use of the flag tool in some artworks where the reason is use of AI. The problem is that I see 99% of the artworks flagged for this reason, in my opinion, have no flaws or anything else that could indicate AI use. The only thing they have in common is the artist having an art style similar to AI artworks, although this is merely the artist's art style, and most have been drawing long before this AI craze, a fact that was completely ignored by the user who flagged it.

Basically, I see a somewhat malicious use of the flag tool in this regard. They're basically flagging based solely on the art style and completely ignoring the lack of evidence of AI use or even signs the art is hand-made. I even strongly suspect these flags are being made by the same user, or at most two, as the pattern is very similar and the reason for them doing this is unknown.

Based on this, I would like to propose changes regarding the checking of artworks suspected of using AI. The first would be the creation of a pinned thread here on the forum, "Is this AI?", where users who merely suspect that an artwork is created or assisted by AI, but have no solid evidence, could simply submit it to the thread, and we users could analyze and discuss among ourselves and the site staff whether or not AI was used. This thread could even help janitors make decisions about whether or not to delete an artwork suspected of using AI, based on what users discuss in this thread. This is 1000 times better than immediately flagging it and having a long debate in the artwork's comments, with people thinking it is and others thinking it isn't. This debate right on the comments could cause distress and depression for the artist, as I've seen happen.

Another point of change is that if a user decides to immediately flag an artwork due to AI use, they must submit solid evidence of AI use, pointing out flaws that could indicate AI use, or even statements from the artist "confessing" to having used AI. Flagging without submitting evidence or based solely on "vibes" could result in the user being punished for abusing the flag tool.

A public thread could be a ripe cesspool of artist witch hunting. That's why it's not conductive to post accusations of generative AI use in public comments. I don't think having a public forum for this would be a good idea.

A topic for this would definitely result in centralized witch hunts, these kinds of topics are explicitly discouraged and often locked
Flags are meant to be used when someone suspects something is wrong with a post, or in other words, they're being used correctly

Even if the flag isn't exactly detailed the individual user should be redirected rather than trying to dump everyone into the forums, a feature less than 1% of the site knows exists
Not to mention that the forums are crawling with power users that would tear users that make accusations against their favorite artists apart limb from limb

I don't think we need this.

as mentioned above, a big reason why such a thread shouldn't exist is because it'll start witch hunts from people who think they've found AI art, or who agree that someone else found AI art. in a vacuum, if everyone could behave rationally and only tried to use facts to try and prove or disprove AI imagery, then a public thread could help but not all people will have good intentions when bringing posts to a public thread, and not all people will know what to look for or will simply agree because someone else "sounds smart, therefore they must be right". the risk of people getting riled up towards innocent images and artists is too great in contrast to having users flag it, with the downsides to flags being it may take longer and it requires manual work...

I think any method of bringing posts to janitorial scrutiny that doesn't involve flagging them with the Flag link is a complete waste of everybody's time.

Staff are currently discussing how to potentially improve how we handle flagging & investigating AI works; however, something like this is so far off of the table it's added another crater on the moon, for the previously stated reasons. A reminder that us mods still have to clean up AI accusation comments (PSA, that's against the rules & you can get records for it; make a flag & carry on with your day), & I foresee creating a dedicated place for that being seen as tacit permission instead of quarantine.

If you're worried about someone abusively creating false flags, we can see their flags; janitors would start recognizing the flagger & check their profile, where there's an extra panel (similar to comments & forum posts) listing how many flags they have. Going to that linked page would make it very easy to spot abuse.

If you have concerns about someone specific, file a report for the user & us mods can look into it.

An idea I have to help aid against flag abuse on the basis that I'm assuming they are being made by newcomers who don't know the rules is a feature for "user moderation level"

I have an account on another site that also lets users flag and report but user also have the ability to remove uploads

I'm not saying to give us user here the ability to remove but rather the user moderation level

When a new account is made you could upload like normal but depending how long you had the account to learn how things go you eventually gain the ability to flag posts but instead of being able to flag for anything you can only flag for uploading guidelines. Then after more time and how true your flags and reports were you gain the ability to flag for other things like AI generated then young human and so on and so forth

We already have user that have privilege status for uploading and edits maybe we can have an in-between like member+ based on the age of your account

It's and Idea and It might not work for those that already have an old account and are probably flagging due to AI paranoia and their disgust for cub content but staff could still see their flag history like Aacafah pointed out

fliphook said:
(a lot of stuff)

Restricting flagging will just push people to create tickets instead, further making the problem worse by putting it in the wrong queue (this regularly happened before the restrictions on the uploading guidelines flag were loosened)

The uploading guidelines flag is also the ai flag, so in the end this would do nothing

donovan_dmc said:

The uploading guidelines flag is also the ai flag, so in the end this would do nothing

How often are tickets used and can't the flag list be reworked for more specific flags

fliphook said:
How often are tickets used and can't the flag list be reworked for more specific flags

tickets are used constantly, and that ends up being burdensome because moderators are not equipped to handle post deletions, flagging, and unflagging.
It also adds more onto their ticket queue.

also you can add reasoning for flags again so I don't see the need for specific flags, other than just drawing more attention, which can leave legit artists whose works were falsely flagged as ai, being bullied or have their names dragged through the mud just because some user doesn't know how to count toes.

I don't think there's a good reason to have flags visible and searchable to standard users. It draws weird attention to images from people looking for drama. I'm sure it'd be possible for repeat flags to be blocked from being made.

regsmutt said:
I don't think there's a good reason to have flags visible and searchable to standard users. It draws weird attention to images from people looking for drama. I'm sure it'd be possible for repeat flags to be blocked from being made.

yes. we can lock flags if they're abused, recently had to do this with a cub piece that had gotten flagged like 4 times in a week or two

Honestly, and this is just personal opinion (so take it with a grain of salt), if you are certain, 100% without a doubt certain that a piece is AI, flag it. Otherwise, if I am not certain it is or not, what I'd do, is ask a janitor to look it over.

Just a quick little poke, ask them, "Hey, do you think this is AI? I'm not sure if I should flag it." From there, they can give you a better understanding of what to do.

Creating a public forum of potential bad actors only leads to witch hunts against anyone that doesn't perfectly fit the mold, innocent or otherwise. Learned that one from experience.

regsmutt said:
I don't think there's a good reason to have flags visible and searchable to standard users. It draws weird attention to images from people looking for drama. I'm sure it'd be possible for repeat flags to be blocked from being made.

Them being visible prevents duplicates (especially since without some ui gymnastics it won't let you make another one while one is active)

It also was changed and while back so only the most recent flag is shown to regular users, they can't see the history

donovan_dmc said:
Them being visible prevents duplicates (especially since without some ui gymnastics it won't let you make another one while one is active)

It also was changed and while back so only the most recent flag is shown to regular users, they can't see the history

Yeah, I was more thinking along the lines of either ui changes or just 'if this flag already exists another can't be made' that would be invisible to users.

bird-tm said:
Honestly, and this is just personal opinion (so take it with a grain of salt), if you are certain, 100% without a doubt certain that a piece is AI, flag it. Otherwise, if I am not certain it is or not, what I'd do, is ask a janitor to look it over.

Just a quick little poke, ask them, "Hey, do you think this is AI? I'm not sure if I should flag it." From there, they can give you a better understanding of what to do.

From earlier in the thread:

lafcadio said:
I think any method of bringing posts to janitorial scrutiny that doesn't involve flagging them with the Flag link is a complete waste of everybody's time.

Though that is just his opinion

snpthecat said:

bird-tm said:
Honestly, and this is just personal opinion (so take it with a grain of salt), if you are certain, 100% without a doubt certain that a piece is AI, flag it. Otherwise, if I am not certain it is or not, what I'd do, is ask a janitor to look it over.

Just a quick little poke, ask them, "Hey, do you think this is AI? I'm not sure if I should flag it." From there, they can give you a better understanding of what to do.

From earlier in the thread:

lafcadio said:
I think any method of bringing posts to janitorial scrutiny that doesn't involve flagging them with the Flag link is a complete waste of everybody's time.

Though that is just his opinion

Sending a message to a single janitor will almost certainly result in the same outcome a flag would have, it'll be brought up and discussed with other Janitors
But with messaging someone directly you run the risk of them not seeing the message, where as that is not a risk with a flag
Even when I was staff I got those messages despite not being a janitor, they just got forwarded on to the rest of the staff like a regular flag would have

Just flag the post and let the staff deal with the people that complain, they're well used to it by now

Additionally, messaging a janitor puts the onus of handling it on them, instead of leaving it to either the most capable/willing janitor or the janitor team as a whole.

It also creates the expectation of a reply, which is just more for them to worry about.

It's also pointless if somebody else just makes a flag for it in the interim. This happens a lot with people reporting posts for violating the Uploading Guidelines instead of flagging them (despite the UI pretty robustly pushing them not to); by the time we see it & get to it, it's usually already been flagged (& sometimes resolved), & even when it hasn't, most mods can't approve/reject posts, so we either have to justify nagging the janitors to handle it or tell them "Can't help you, please flag it instead".

The system exists for a reason. If we wanted to change user behavior, we'd change the system.

This is such a hard one and a good example of what sounds like a good idea could be terrible in practice.

I have fallen for at least one artist at a glance, however they put it in their profile. Had they not and they already had a bunch of art here, I could have made a mistake. I guess searching artist name + status:deleted is one way to check for a history of AI stuff posted and deleted here, but that's far from perfect.

I'm not sure if there's an easy way to search for deletion reasons or if it's standard to tag ai_generated on posts deleted for being AI generated. If so then searching "status:deleted ai_generated" could help.

I think this site needs to relax the rules on AI works. Full AI works should remained banned, but AI assistance should be allowed. Most new artists I know are using AI somewhere in what they do. It's just going to keep growing and expanding.

It would also stop the witch hunts.

freebatforhire said:
I think this site needs to relax the rules on AI works. Full AI works should remained banned, but AI assistance should be allowed. Most new artists I know are using AI somewhere in what they do. It's just going to keep growing and expanding.

It would also stop the witch hunts.

A site's rules do not necessarily counteract the culture and feelings of the people using it. Things such as cub and feral are contentious but are allowed on this website, with many artists and users disapproving of their allowal, and genAI would be the same way. Even if it was allowed, there would be a significant amount of disdain for it from the userbase. Witch hunts happen because AI is seen as bad by a lot of artists/active users on this website, not because they're disallowed. They happen all the time outside of e621 to perfectly innocent artists making things entirely by hand.
That, and artists have gone DNP due to e6AI existing and e621 being scraped for genAI use against the artists' wills. It's a lot more complex of a problem than you'd think it is, but for now e6AI acts as a quarantine to at least keep the pro-AI people happy.

zeorp said:
This is such a hard one and a good example of what sounds like a good idea could be terrible in practice.

I have fallen for at least one artist at a glance, however they put it in their profile. Had they not and they already had a bunch of art here, I could have made a mistake. I guess searching artist name + status:deleted is one way to check for a history of AI stuff posted and deleted here, but that's far from perfect.

I'm not sure if there's an easy way to search for deletion reasons or if it's standard to tag ai_generated on posts deleted for being AI generated. If so then searching "status:deleted ai_generated" could help.

delreason:*ai*generated*

moonlit-comet said:
A site's rules do not necessarily counteract the culture and feelings of the people using it. Things such as cub and feral are contentious but are allowed on this website, with many artists and users disapproving of their allowal, and genAI would be the same way. Even if it was allowed, there would be a significant amount of disdain for it from the userbase. Witch hunts happen because AI is seen as bad by a lot of artists/active users on this website, not because they're disallowed. They happen all the time outside of e621 to perfectly innocent artists making things entirely by hand.
That, and artists have gone DNP due to e6AI existing and e621 being scraped for genAI use against the artists' wills. It's a lot more complex of a problem than you'd think it is, but for now e6AI acts as a quarantine to at least keep the pro-AI people happy.

My point wasn't about keeping people happy. Most new artist, at least that I know and follow, are using AI as part of their process. Eventually e6 is only going to have old artist. Maybe a small trickle of new ones who are extremely passionate.

If some AI is allowed, witch hunts become pointless because it's not against the rules, and artists don't have to jump through hoops to prove their not using AI. Seriously, being forced to film yourself drawing is nuts. Whatever, there are other sites out there.

freebatforhire said:
I think this site needs to relax the rules on AI works. Full AI works should remained banned, but AI assistance should be allowed. Most new artists I know are using AI somewhere in what they do. It's just going to keep growing and expanding.

It would also stop the witch hunts.

AI assisted work already WAS allowed previously, but that created debates of what degree of assistance was allowed. It's easier for enforcement and consistency to blanket ban it rather than debate if someone's mostly-ai linework is okay.

Maybe eventually ai will be ubiquitous and undetectable enough to make rules against it futile. Until then go to e6ai or any number of other websites that allow it.

regsmutt said:
AI assisted work already WAS allowed previously, but that created debates of what degree of assistance was allowed. It's easier for enforcement and consistency to blanket ban it rather than debate if someone's mostly-ai linework is okay.

Maybe eventually ai will be ubiquitous and undetectable enough to make rules against it futile. Until then go to e6ai or any number of other websites that allow it.

Full paintovers specifically were allowed, and backgrounds are still allowed

freebatforhire said:
It would also stop the witch hunts.

freebatforhire said:
If some AI is allowed, witch hunts become pointless because it's not against the rules, and artists don't have to jump through hoops to prove their not using AI.

I think you greatly misjudge how hated AI is in the art space, we previously allowed full paintovers and artists still got ripped to shreds over that

freebatforhire said:
Seriously, being forced to film yourself drawing is nuts.

No one is being forced to fully film their drawing process
The only thing needed to prove it is layers and/or WIPs, the same things asked for to disprove tracing accusations

freebatforhire said:
Whatever, there are other sites out there.

Yes there are other sites out that which will happily welcome all forms of ai slop, we are not one of them

freebatforhire said:
I think this site needs to relax the rules on AI works. Full AI works should remained banned, but AI assistance should be allowed. Most new artists I know are using AI somewhere in what they do. It's just going to keep growing and expanding.

It would also stop the witch hunts.

If relaxing policy had any effect on "witch hunts", then we would never see AI accusations on Twitter because Twitter has always allowed AI-generated images. In reality, it is one of the most common places for people to throw out accusations.
I think you don't understand the userbase and you are just personally pushing what you want to see.

Additionally, not being against the rules has done little to help people erupting over properly tagging young posts, or false-flagging contentious content. Just like that, we'd still force it to be tagged correctly, and we'd still have to adjudicate it. While I agree that people's opinions can be shaped by rules/laws/conventions, it's not instantaneous, nor guaranteed, nor absolute.

freebatforhire said:
...and artists don't have to jump through hoops to prove their not using AI. Seriously, being forced to film yourself drawing is nuts.

We're not punting works from the site because some random user (or some random site) says it's AI; artists only have to actively prove it when there are multiple, significant indicators that AI was utilized, & it's very rare that we'd require any form of video evidence for proof. The only times we ask for it is as a last-ditch effort to give someone an opportunity to prove us wrong because every single solitary piece of evidence was wildly insufficient or actually supported the conclusion they use/used AI. Speedrunners don't ask runners play with a handcam unless they already know they're cheating, & we don't ask artists to submit video of their process unless we already know they're using AI; it's us doing our due diligence to give them every chance we can to prove us wrong.

freebatforhire said:
Whatever, there are other sites out there.

There sure is, & they're more than welcome to it. We don't want it here, for the foreseeable future at the very least. Period. End of sentence. This isn't up for discussion.

snpthecat said:
delreason:*ai*generated*

Thank you very much. This works as an alternative to check artist history without outright posting a wall of shame type list.