Posted under Art Talk
This topic has been locked.
I was browsing my profile and noticed that a Lucario image created by that artist was missing. I checked the comments section and it said it was removed because the artist used AI to create them.
Is that true? I need more explanation.
Updated by Aacafah
zephydx said:
I was browsing my profile and noticed that a Lucario image created by that artist was missing. I checked the comments section and it said it was removed because the artist used AI to create them.Is that true? I need more explanation.
I've followed this artist for years, i really believe (and hope) their work aren't AI Generated.
Maybe a false flag? I've seen lots of people doing it with the "Young Human" thingy
pervertamura said:
I've followed this artist for years, i really believe (and hope) their work aren't AI Generated.Maybe a false flag? I've seen lots of people doing it with the "Young Human" thingy
That's just the thing. Recently found out here that it isn't always a nothing or everything.
They could have been making art by hand up until this point. Not saying that's the case but,
It could be like Speedrunners cheating. They don't do it because they can't, They do it so they
can get there faster. It's a shame for the peeps who fall for that, Drawing isn't the yata yata,
It's the reason to make art, Dood = =)
notkastar said:
That's just the thing. Recently found out here that it isn't always a nothing or everything.
They could have been making art by hand up until this point. Not saying that's the case but,
It could be like Speedrunners cheating. They don't do it because they can't, They do it so they
can get there faster. It's a shame for the peeps who fall for that, Drawing isn't the yata yata,
It's the reason to make art, Dood = =)
Honestly, i have zero ideas on how to respond on this properly, but i can see and understand your point, despite not agreeing it fits the current situation/context, but again, i can see your point... Sadly
Not just the Lucario piece, but practically everything ever uploaded: azoomer status:any. Certainly not a false flag, but a very intentional decision.
I assume we'll never see what evidence the staff used to make this decision, but that certainly is quite an accusation. aZoomer seems to deny it but I do see a sudden change in their style circa 2023.
I've always been a little suspicious of the style, but assumed an artist with such a following would have been exposed much sooner if it was the case.
faucet said:
Not just the Lucario piece, but practically everything ever uploaded: azoomer status:any. Certainly not a false flag, but a very intentional decision.I assume we'll never see what evidence the staff used to make this decision, but that certainly is quite an accusation. aZoomer seems to deny it but I do see a sudden change in their style circa 2023.
I've always been a little suspicious of the style, but assumed an artist with such a following would have been exposed much sooner if it was the case.
Same.... It's really sad... But oh well, hope is the last one to die.
Updated
faucet said:
I've always been a little suspicious of the style, but assumed an artist with such a following would have been exposed much sooner if it was the case.
yeah, looking at the artist's FA, there's a distinct, sudden shift in the art style. about half of the way down on page 3 of their gallery the art above and below a line seems like they're drawn by two entirely diffrent artists...
which, benefit of the doubt, it's not an impossibly that an artist shifts their style, especially relatively early on in their career (the first post in the gallery is only two years prior to the shift). I've certainly seen that kinda thing before.
Quite honestly this needs an explanation from the mods, this is incredibly ridiculous. To me this look like a mod gone rogue.
Chimi/Azoomer is very good at what they do, they have posted plenty of their own tutorials and speedpaints that show their process. I am not sure why E6 would completely remove all posts from a legitimate artist and accuse them of using AI.
I'd welcome any actual evidence. Right now this decision is slandering the reputation of a well-known and reputable artist. And there does not appear to be any legitimate evidence supporting it.
That's a very disappointing false flag, I would push for mods to re-instate this artwork since it's human-created after all.
Azoomer quote: "Do i have to show every single timelapse if i ever want to upload on [e621] site or what? [e621] even deleted a video i showed the entire process before, i had uploaded this for patreons just a few days ago."
And sure enough, I do see timelapses of Azoomer creating their artwork. If mods are deleting *those as well* then this goes beyond the pale.
pm_me_hot_eevee said:
Quite honestly this needs an explanation from the mods, this is incredibly ridiculous. To me this look like a mod gone rogue.Chimi/Azoomer is very good at what they do, they have posted plenty of their own tutorials and speedpaints that show their process. I am not sure why E6 would completely remove all posts from a legitimate artist and accuse them of using AI.
I'd welcome any actual evidence. Right now this decision is slandering the reputation of a well-known and reputable artist. And there does not appear to be any legitimate evidence supporting it.
Hi, rogue blade runner janitor mod here. My peers are already in the loop about these deletions, and they know exactly (down to specific dates and images) what I consider to be suspicious about aZoomer's portfolio. Which is to say, I am acting with full knowledge of a direct superior who'd be able to shut this down instantly if I didn't have a reasonable case.
Depending on the submitted evidence for this case we either may or may not make a public statement later. This is dependent on the final result after appeals are resolved, what parts of the information are suitable for public release, etc.
That would explain why the art kept giving me AI vibes.
Well there's $100 down the drain..
Seeing retribution for sloppers is always so satisfying.
afterglow said:
That's a very disappointing false flag, I would push for mods to re-instate this artwork since it's human-created after all.Azoomer quote: "Do i have to show every single timelapse if i ever want to upload on [e621] site or what? [e621] even deleted a video i showed the entire process before, i had uploaded this for patreons just a few days ago."
And sure enough, I do see timelapses of Azoomer creating their artwork. If mods are deleting *those as well* then this goes beyond the pale.
Tracing AI is a thing that people can do, you can very easily hide layers from timelapses in at least few popular art programs, and there are programs that allow overlaying images over art programs without it showing in timelapses even if you use program that do not allow layer hiding. So timelapses alone are not a bulletproof evidence that AI was not used.
afterglow said:
That's a very disappointing false flag, I would push for mods to re-instate this artwork since it's human-created after all.Azoomer quote: "Do i have to show every single timelapse if i ever want to upload on [e621] site or what? [e621] even deleted a video i showed the entire process before, i had uploaded this for patreons just a few days ago."
And sure enough, I do see timelapses of Azoomer creating their artwork. If mods are deleting *those as well* then this goes beyond the pale.
Those timelapses are incredibly obvious as AI to anyone with decent art experience, as for other stuff he's also traced from real art as well, not just AI, and the few times he's gone live on stream its just to "study" by copying someone else's work, its a consistent pattern of behavior.
Of all the AI tracers out there, azoomer has been a known asset the longest. This drama came up two years ago and they managed to survive it because of the size of their audience. I would also say they're one of the worst offenders because not only have they traced AI, they do so with contempt for the craft of art. They have logs leaked - earlier in this thread, and elsewhere - where they specifically test their art against AI detectors (these are useless btw) and joke about how "nobody will ever know".
When the drama first arose they also made a statement explicitly admitting to using AI, though they now claim that they only ever used it for two images. This is a fabrication because you can clearly see in their FurAffinity gallery the point at which they started using AI and their art, save for one or two pieces, never returned to their pre-AI skill level which they were posting only a month prior.
They continue to lie about the degree to which they used AI in their process, have never owned up to the to their duplicitous comments about tricking viewers, and never disclosed to commissioners that they had been or could be using AI. I have several friends who got commissions from them who would never have done so had they known azoomer ever used AI at all.
While I do generally respect that it's possible they don't trace anymore, they undeniably built their entire artistic career off tracing and we have no way to determine how many commissions they've done that were traced nor can we ever be sure they aren't still tracing especially given the contempt they've shown in private for the craft of art. They've also been caught tracing real artists so there's that, too. All of these things I'm sure the e621 moderators have seen.
Why it took two years for them to get removed from e621 I'm not sure. For my part, I've reported them to FurAffinity several months ago but FA is *very* strict about proof, to the point that they might as well not even ban for it, as not only did they not ban azoomer they didn't even remove the art pieces in their gallery they specifically admitted were AI-assisted
I know this is a rude awakening for some viewers but we're currently in an era where AI trace grift is at an all-time high. I and several other artists have collectively identified at least 40 persons, some still actively posting here on e621, who are almost certainly using or tracing AI. I prefer not to do call-outs unless there is exceedingly undeniable proof of malfeasance and, unfortunately, those kinds of smoking guns are hard to come by. Of the names we've identified only five have ever had irrefutable proof come out against them (such as Pawralax, who I reported multiple times out of an abundance of suspicion before they messed up and sent a commissioner a WIP with a genAI filename). Azoomer is one of those five; they've directly admitted to using AI in their process, have provided ample evidence through leaked logs of their methods of skirting around detection, and have failed to provide proof to a level acceptable to real artists that they don't require AI to get their current results.
For these reasons, not only can we not be certain about what amount of art in their gallery was created with AI assistance, we can't be certain if they're being truthful about not using AI. It's unlikely they ever would have come forward about it if their Discord logs weren't leaked. Ethically speaking, even if azoomer said they're no longer tracing, there are hundreds of talented artists out who don't use GenAI in any capacity who are worth your time and attention and who you don't have to wonder if they're trying to trick you.
I'm not an e621 mod or anything, I've just been looking at the collective era of AI tracing we're dealing with now, and though this removal should have happened when the drama first broke, now is as good a time as any. I'm sorry for anyone who enjoyed their art or commissioned them, but you have to remember it's not your fault. The entire goal of these grifts is specifically to trick you, and even real artists sometimes fail to detect AI-assisted artwork. The important thing is that more people know and can make informed decisions about what level of suspicion or trust they're willing to engage with.
It's just so obvious too from a glance when you go back to two FA posts from February 20, 2023 to March 24, 2023
Such a crazy amount of style change and growth, all within a month?
It's not impossible for someone to put in a lot of work in a month long period, but it's pretty uncommon lol
donovan_dmc said:
Well there's $100 down the drain..
Can't you request a refund for art that (apparently) was created using AI? :с
notevilfox said:
Can't you request a refund for art that (apparently) was created using AI? :с
The most likely outcome of that is both of our paypal accounts being suspended since payments for adult content are explicitly against their tos
donovan_dmc said:
The most likely outcome of that is both of our paypal accounts being suspended since payments for adult content are explicitly against their tos
I know of cases where refunds were made and they didn't tell Paypal that the art had adult content... would Azoomer agree to refund your money itself if there is evidence that AI was used?
It's sad to leave this situation just like that 😓
Taking in everything on this situation, I think there is absolutely no way azoomer can be let on. Even IF their last ten images were completely hand drawn and they can prove it with a video camera strapped to their ceiling recording them drawing it raw, there is just far too much fuckery to trust them now.
The big art shift, the many posts from them talking about USING AI... It's a tall glass of water that is far too muddy for me, or anyone to drink down.
Our resident bladerunner janitor probably has much more compelling arguments and evidence already put together though
fuzzikayu said:
When the drama first arose they also made a statement explicitly admitting to using AI, though they now claim that they only ever used it for two images.
Do you happen to have that quote somewhere? That sounds like it'd be really useful here.
notevilfox said:
I know of cases where refunds were made and they didn't tell Paypal that the art had adult content... would Azoomer agree to refund your money itself if there is evidence that AI was used?
It's sad to leave this situation just like that 😓
You could try asking them. If they haven't finished it, maybe they'll give it back. If they DID finish your work.. May be SOL on that front. Only thing I could think to try then is a chargeback, but I've never tried it personally so I can't say much if that would work, and ofc like donovan said that comes with the risk of an account nuke from orbit.
cowboyjones said:
Do you happen to have that quote somewhere? That sounds like it'd be really useful here.
Not Fuzzikayu but here are the ones I got https://imgur.com/a/xusOqJR
Also the last discord post is false. I asked Waga if he really apologized but of course he didn't. Said it was 100% AI.
ruppari said:
Tracing AI is a thing that people can do, you can very easily hide layers from timelapses in at least few popular art programs, and there are programs that allow overlaying images over art programs without it showing in timelapses even if you use program that do not allow layer hiding. So timelapses alone are not a bulletproof evidence that AI was not used.
fuzzikayu said:
Of all the AI tracers out there, azoomer has been a known asset the longest. This drama came up two years ago and they managed to survive it because of the size of their audience. I would also say they're one of the worst offenders because not only have they traced AI, they do so with contempt for the craft of art. They have logs leaked - earlier in this thread, and elsewhere - where they specifically test their art against AI detectors (these are useless btw) and joke about how "nobody will ever know".When the drama first arose they also made a statement explicitly admitting to using AI, though they now claim that they only ever used it for two images. This is a fabrication because you can clearly see in their FurAffinity gallery the point at which they started using AI and their art, save for one or two pieces, never returned to their pre-AI skill level which they were posting only a month prior.
They continue to lie about the degree to which they used AI in their process, have never owned up to the to their duplicitous comments about tricking viewers, and never disclosed to commissioners that they had been or could be using AI. I have several friends who got commissions from them who would never have done so had they known azoomer ever used AI at all.
While I do generally respect that it's possible they don't trace anymore, they undeniably built their entire artistic career off tracing and we have no way to determine how many commissions they've done that were traced nor can we ever be sure they aren't still tracing especially given the contempt they've shown in private for the craft of art. They've also been caught tracing real artists so there's that, too. All of these things I'm sure the e621 moderators have seen.
Why it took two years for them to get removed from e621 I'm not sure. For my part, I've reported them to FurAffinity several months ago but FA is *very* strict about proof, to the point that they might as well not even ban for it, as not only did they not ban azoomer they didn't even remove the art pieces in their gallery they specifically admitted were AI-assisted
I know this is a rude awakening for some viewers but we're currently in an era where AI trace grift is at an all-time high. I and several other artists have collectively identified at least 40 persons, some still actively posting here on e621, who are almost certainly using or tracing AI. I prefer not to do call-outs unless there is exceedingly undeniable proof of malfeasance and, unfortunately, those kinds of smoking guns are hard to come by. Of the names we've identified only five have ever had irrefutable proof come out against them (such as Pawralax, who I reported multiple times out of an abundance of suspicion before they messed up and sent a commissioner a WIP with a genAI filename). Azoomer is one of those five; they've directly admitted to using AI in their process, have provided ample evidence through leaked logs of their methods of skirting around detection, and have failed to provide proof to a level acceptable to real artists that they don't require AI to get their current results.
For these reasons, not only can we not be certain about what amount of art in their gallery was created with AI assistance, we can't be certain if they're being truthful about not using AI. It's unlikely they ever would have come forward about it if their Discord logs weren't leaked. Ethically speaking, even if azoomer said they're no longer tracing, there are hundreds of talented artists out who don't use GenAI in any capacity who are worth your time and attention and who you don't have to wonder if they're trying to trick you.
I'm not an e621 mod or anything, I've just been looking at the collective era of AI tracing we're dealing with now, and though this removal should have happened when the drama first broke, now is as good a time as any. I'm sorry for anyone who enjoyed their art or commissioned them, but you have to remember it's not your fault. The entire goal of these grifts is specifically to trick you, and even real artists sometimes fail to detect AI-assisted artwork. The important thing is that more people know and can make informed decisions about what level of suspicion or trust they're willing to engage with.
This whole thing is so sad, the outcome of such situation is yet to be known, but this isn't the first time it happened, first it was HelicopterFood/Kuru Tyan, now, Azoomer, almost the same situation, i feel for the people affected (such as Commissioners) by those cases.
I can't even imagine a way to avoid this kind of situation, if even screen-recording can't be taken seriously and is being cheated, what an artist can use to prove, body cams?
Sadly, this is only set to continue for a long time.
If you use it to help you, like improving shadows or something like that, then isn't very different from Photoshop or other image editors. I believe if you'll ban someone for using an AI to choose a better color for his drawing, then you should ban half of the content of this website for editing with Photoshop, wich automates processes way before AI being announced. The point is, if the image isn't fully AI generated, it is AI edited, which like I said, isn't nothing different from image editors.
phustrak_the_dragon said:
If you use it to help you, like improving shadows or something like that, then isn't very different from Photoshop or other image editors. I believe if you'll ban someone for using an AI to choose a better color for his drawing, then you should ban half of the content of this website for editing with Photoshop, wich automates processes way before AI being announced. The point is, if the image isn't fully AI generated, it is AI edited, which like I said, isn't nothing different from image editors.
This is such a bad take argument.
First off, if artists want to understand shadow, there's actual books and videos teaching you how to study shadows properly. You can take your own pictures and use those for study if you're trying to become more realistic. Put in the work properly, and you won't need AI at all to learn, which is a benefit since researchers are finding that AI leads to poorer quality learning.
Secondly, AI does not understand shadow works, it just knows "this part is darker in x% of images". Photoshop is not suggesting you shade this portion of the image, it's giving you tools to do the shading yourself. This is unlike AI, which will edit your pic if you tell it to do so, and not in ways that you learn from. That's the entire reason that AI is not like a calculator, another common argument.
AI can be a useful tool for some things, like frequency analysis through noisy data (hey, that's where it came from!). It's not, however, conducive to the type of art E6 archives, it just leads to factory-churned art, which Azoomer directly bragged about in the linked Discord messages above.
phustrak_the_dragon said:
If you use it to help you, like improving shadows or something like that, then isn't very different from Photoshop or other image editors. I believe if you'll ban someone for using an AI to choose a better color for his drawing, then you should ban half of the content of this website for editing with Photoshop, wich automates processes way before AI being announced. The point is, if the image isn't fully AI generated, it is AI edited, which like I said, isn't nothing different from image editors.
How is this relevant to the thread? aZoomer's works were deleted because they were traced from AI generated images. Even if AI wasn't involved their work would still be deleted for being traced.
phustrak_the_dragon said:
If you use it to help you, like improving shadows or something like that, then isn't very different from Photoshop or other image editors. I believe if you'll ban someone for using an AI to choose a better color for his drawing, then you should ban half of the content of this website for editing with Photoshop, wich automates processes way before AI being announced. The point is, if the image isn't fully AI generated, it is AI edited, which like I said, isn't nothing different from image editors.
There are rules for AI related content as per guidelines https://e621.net/wiki_pages/11143
Trying to upload artwork that was generated by AI or heavily assisted by it (in this case, tracing) isn't allowed. If you want to upload such content, there is a sister site e6ai for that.
There's also an unspoken rule that you should rather disclose that you are using AI in any form to your consumers if they are paying for it. Trying to pass it off as non-AI is just shameful. There should be honesty and transparency, not trickery and obfuscation.
Just like tracing other people's artwork is already unethical if you do not disclose it, so would be using AI generative or heavily assisted content.
lafcadio said:
Hi, rogueblade runnerjanitormod here. My peers are already in the loop about these deletions, and they know exactly (down to specific dates and images) what I consider to be suspicious about aZoomer's portfolio. Which is to say, I am acting with full knowledge of a direct superior who'd be able to shut this down instantly if I didn't have a reasonable case.Depending on the submitted evidence for this case we either may or may not make a public statement later. This is dependent on the final result after appeals are resolved, what parts of the information are suitable for public release, etc.
I'd be curious to know how the blade runners came to that conclusion. I look at AI genned porn all day as an e6ai janitor and can say the art is way far removed from the usual slop I see.
phustrak_the_dragon said:
If you use it to help you, like improving shadows or something like that, then isn't very different from Photoshop or other image editors. I believe if you'll ban someone for using an AI to choose a better color for his drawing, then you should ban half of the content of this website for editing with Photoshop, wich automates processes way before AI being announced. The point is, if the image isn't fully AI generated, it is AI edited, which like I said, isn't nothing different from image editors.
That is not what they did lmao, you're just trying to find a justification for genAI so you can feel better about it and yourself.
If someone is curious, they're welcome to DM the deleting janitor to discuss the matter, but as explained in multiple threads, we are not in the habit of publicly discussing our detection methodologies for multiple reasons, including (but not limited to):
We encourage artists of posts that were incorrectly determined to use disallowed AI generation/assistance to appeal the matter to us; if an artist can satisfactorily demonstrate an ability to create the accused works without disallowed AI generation/assistance (& can sufficiently discredit any evidence that a specific work uses disallowed AI generation/assistance, should such evidence exist), we've no problem with admitting we were wrong, and of course immediately & fully restoring the accused posts.
As Lafcadio said, in some select situations, we do create write-ups discussing our findings, but not only are these exceptionally rare cases where the need to put the matter to bed permanently arises, to my knowledge we do not make these without the accused first coming forward to appeal the matter, and only in certain circumstances (e.g. they are completely incapable of providing any substantive evidence to discredit the verdict) do we feel confident enough in our verdict to make such definitive presentations of our verdict.
As for the assertion that regularly reviewing openly AI generated works would lend credibility, I'd agree that it would hold more weight than a lay user, but this is still a far cry from a trained eye looking for duplicative AI assistance/generation. Especially when an individual has some genuine, unassisted, & wholly organic artistic ability, it is trivially easy to create works that are exceptionally difficult to accurately detect, even to an eye trained in analyzing works they know use AI assistance/generation, and the ability to evade detection increases dramatically with the raw ability of the artist. As I've said previously:
It's kinda like speedrunners; someone might have the knowledge & skills to do something legitimately, but that also means they have the knowledge & skills to take a shortcut and hide it better than someone less capable, along with an emotional ("I know I can do it, but this is easier, & besides, I've earned this") and/or financial motivation to do so.
To continue the parallel, while individuals well-versed in creating or picking out edited or manipulated gameplay (e.g. video editors, mod developers, other cheaters) may have an edge in detection over the average viewer, it's a far cry from the detection capabilities of an experienced speedrunner of that game. A video editor could detect errors in the audio/video splicing, inconsistent audio mastering, or even an alleged unedited console capture that cannot be authentically generated by the system, but you'd need a runner's experience to even think to look for indicators like:
Obviously, this is an analogy & not perfectly 1 to 1 with AI art, but I hope this illustrates how familiarity with the process alleged doesn't mean you'll have substantially better detection abilities than someone with familiarity with the actual underlying activity, nor someone specifically looking for something inauthentic in a wider context.
For a more direct example, say you have a talented artist that has a large gallery of wholly unassisted, impeccable, & extravagantly detailed works, with extremely skilled texturing & shading... but it takes a lot of time & effort. Maybe they get tired of it, maybe their finances aren't where they want them to be, or maybe they don't have the time to keep putting this much into these work, but whatever the reason, say they decide to just do the linework & flat coloration, and use that as input on a small, local model trained on their work specifically to get the bulk of the shading work done, & they just refine the output directly, the same way they would if they did it themselves. How would a keen eye for AI errors help you detect this if they do a good job at catching them before publication, & they have the skills to manually correct them? That's not to say such a skillset would provide no added advantage in this case, but it is still miles from a broader analytical approach that takes into account the artist's previous works and how they might differ from their later, more suspect works.
I hope this has assuaged any concerns on the matter. If the artist would like to contest, they have a path to do so, but we are not fielding public inquiry on the matter at this time.
