Derpibooru
Filters (Default) SettingsRegisterLogin

Regarding AI-generated Imagery

Site and Policy » Regarding AI-generated Imagery Search Posts
Started by Princess Celestia
Posted
76 replies
Subscribe
Background Pony #38F5
@CookieGalaxy
Oh, I see now. I thought you just meant a single tag called ai-director, not a new prefix for tags.
Well, whether that becomes a thing or if the site sticks with artist: as the main prefix to identify the primary human action taker (which you could easily complex-filter as artist:whoever, machine learning generated), I agree that the people should be indicated, not omitted.
Background Pony #38F5
@CookieGalaxy
And author:, and colorist:, and commissioner: - but all except artist: mean someone else was involved: whatever was colored or edited, or that the subject of interest in the photograph is someone else’s craft like a cosplay or plushie, or that it’s a comic that one person wrote and another person drew. This is the basis by which I feel that artist: is used for specifying primary responsibility, at present, and to think hard about whether to stray from that.
It’s even used for tagging the person responsible, for example, of playing the game Minecraft to lay blocks as pixels, even if they didn’t make the pixel art design that’s recreated in the game (that other person may also be tagged as an additional artist:, if identified). (Trying very hard not to descend into “whataboutism” - just trying to demonstrate the breadth of the artist: prefix.)
Posted Report
Background Pony #ECE7
Can you add style artist: tag for AI-generated art where an artist name was specified specifically to imitate their style? There needs to be some way to give credit back to the people that created a style, and there’s currently no way to do it.
Posted Report
Background Pony #ECE7
@Parallel Black
Okay. But people are going to do it, and I don’t control everyone. When people inevitably do it, should they have a way to credit the original style artists or not? I think they should.
Posted Report
Background Pony #ECE7
@PUBLIQclopAccountant
It should work for both human and AI images. If I like some artist’s style and decide to create something strongly reminiscent of it, it would be good for them to get credit whether or not I used an AI to imitate their art. It’s good for the original artists so they get attention for their creativity, it’s good for viewers so they know how to find more similar work, and it’s good for the new artists using other’s style since the accurate tagging makes it easier for people to discover their work.
Posted Report
Background Pony #38F5
@Background Pony #ECE7, @PUBLIQclopAccountant, @Background Pony #ECE7
That’s an interesting idea. There are already ~4250 posts tagged style emulation, ~550 tagged fine art parody (and another ~50 tagged fine art emulation), plus another ~1300 among the several I can’t believe it’s not * tags.
Although, I’m not sure whether it’s more appropriate to convey that information via specific tags, or to just write in the description.
Some tricky points that already come to mind:
  • The “I can’t believe it’s not” tags get used also when the style resemblance isn’t intentional. And, consider the reverse, where someone meant to emulate a style but didn’t do so good a job of it - would style tags be used based on intention or on result, and if the latter, how would the judgement be defined? (The “I can’t believe it’s not” tags feel a lot more lighthearted to me and don’t seem troublesome to leave unformalized.)
  • What you get from AI image generators may or may not have any resemblance to speak of to the artists used in a prompt:
    • The influence on the image might be as little as “the named artist tends to work in grayscale, so the outputs will be grayscale”, for example. Or, “the named artist tends to make paintings with fine detail and shading, so the outputs will be detailed and shaded, not flat-colored”. It definitely wouldn’t be appropriate to attribute style to every name in a prompt.
    • And, depending on the generator used, it might end up emulating an artist that isn’t named at all, if that artist is heavily associated with something else - like if you ask for a cubist picture, you might end up with Pablo Picasso’s style.
Posted Report
Background Pony #ECE7
@Background Pony #38F5
I believe specific tags would be better for several reasons:
  • If there are many images that imitate specific styles, such information should be easy for viewers to navigate.
  • It should be easy for artists to show off all images based on their style so they can showcase their creative impact. This is only possible if it’s easy for artists to find derived work based on their style.
  • I don’t believe viewers are substantially different from people generating images with AI. If people want to generate images of a particular style using AI, then it makes sense that viewers will want to navigate images by style. This is only possible if image styles are tagged appropriately.
I believe these new “style artists:” tags should only reflect intention. I say this for two reasons.
  • The point is to make the chain of creative contribution more clear. I believe this is a much more important point to address than the classification of styles, which is what style similarity tags would accomplish.
  • Whether two styles are actually similar is a judgement call, and that involves a very fuzzy line. Tags should be reasonably black-and-white.
Posted Report
Exedrus
Lunar Supporter - Helped forge New Lunar Republic's freedom in the face of the Solar Empire's oppressive tyrannical regime (April Fools 2023).
Best Artist - Providing quality, Derpibooru-exclusive artwork
Princess of Love - Extra special version for those who participated in the Canterlot Wedding 10th anniversary event by contributing art.
Silver Supporter - Silver Supporter
Heart Gem -
Not a Llama - Happy April Fools Day!
Friendship, Art, and Magic (2017) - Celebrated Derpibooru's five year anniversary with friends.
Happy Derpy! -
Responsible Disclosure -
Artist -

@Bigcheese
That was a good read on the subject. Though it doesn’t seem to have many cut-and-dry rulings (for the US at least; I mostly skimmed the EU parts). I do want to point out that your summary is very high level. In particular, this assertion has a lot of caveats:
… Nobody is even currently considering the training set authors as potential copyright owners.
The paper lists a lot of things that could cause that to not be true, such as:
  • If the resulting image looks very similar to a copyrighted work, then it’s probably considered derivative regardless of the AI that was used to generate it. (Though AIs usually try to avoid this.)
  • If an AI is used to run an artist out of business by training off of their works, it is likely the AI’s creation violated the artist’s copyrights.
  • The paper’s authors were unsure whether copyright might apply if an AI copied “an authors’ copyrighted expressions (Sobel, 2017), learning from their creative and expressive choices (Bonadio and McDonagh, 2020).” That is, in many ways, the big question when it comes to these works: if a machine emulates a specific artist’s style, especially if the AI is framed as a tool trained off of those works, are they derivative?
If the work clears those legal hurdles, then it at least hasn’t violated copyright law in its creation. At that point, there’s some guidance on what could merit a copyright:
  • Images generated at random (with no human input for the specific generated images) have no copyright, since no human was really involved in the creative process.
  • Images where a substantial amount of creativity comes from the individual (i.e. the tool is like a really fancy Photoshop brush), will likely grant copyright to the individual.
  • If a bunch of images are generated, then hand-filtered by a human, that might be enough “creative input” to grant a copyright.
The above leaves a tonne of legal grey area, so I’m guessing we’ll be seeing legislatures or supreme courts weighing in in the near future.
Posted Report
🐴
Magnificent Metadata Maniac - #1 Assistant
Solar Guardian - Refused to surrender in the face of the Lunar rebellion and showed utmost loyalty to the Solar Empire (April Fools 2023).
Non-Fungible Trixie -
Magical Inkwell - Wrote MLP fanfiction consisting of at least around 1.5k words, and has a verified link to the platform of their choice

IRL 🎠 stallion
@Exedrus
Typing the first half of this up for the second time (and probably forgetting my overall argument intentions) because the programmers for either iOS and/or iPhone Firefox are defective and lost my half-typed comment when I took a picture of a spiderweb.
To my understanding, it’s only a derivative work if it’s recognizably similar to an existing work. If it’s only vaguely “in the style of”, it’s probably not derivative (unless it happened to copy a watermark as part of the style). Think of the humans-only analogy of hiring an artist to draw some fetish clop “in the style of” a strictly SFW artist. Unless it’s modeled as a trace/edit of an existing drawing by the SFW artist, it’s not getting the edit or traced tags.
That all said, I would not put it past US Congress to totally blow it when it comes to assigning copyright for AI-generated art. The correct answer is that it all should be exempt from copyright except where substantial human transformation was involved (think collages, heavy edits, using the AI image as a base for tracing, or feeding your own art into the AI and asking the computer to redraw it). Sadly, they’ll probably pick the idiotic option of letting overall style be protected IP and turning the art world into the wasteland of frivolous stylistic similarity inanity that plagues the music industry.
Posted Report
Hazy Skies

Aura Dust (AURAequine)
So I had just found out that editor: is a thing now.
I feel this could be problematic as it could essentially double the amount of tags (comparatively to just artist:). In addition, could cause confusion due to the ambiguity of quality AI generated art. Involving Editor: implies editing, which may or may not be accurate. Only the poster, providing they’re the creator, would be aware of the level of work involved.
Another issue with this is affecting existing artist tags. I’m certain there’s others like myself who has both AI and Standard artwork, and would prefer my work not to be mixed up, or have to include multiple extra tags to each (again, depending on the level of work involved.)
Personally, I believe that those using AI are still artists, though of a different kind; prompt artists if you will. Much like regular artists, having many varying levels of skill and technique involved with their craft. It may be considered directing as others have mentioned, but as the AI’s focus is art and is a tool for the end user, much like other art programs are to create art as the end product, then the director should also be considered an artist as well.
As such I think that sticking to just the one tag artist: should suffice. AI artwork are already filtered out from the default filter, there shouldn’t be any further need to isolate the specific artwork other than what’s already defined (machine learning generated)
Thank you for reading, and please consider.
Kind regards,
Hazy Skies
Posted Report
Edited
Background Pony #9FE1
@Hazy Skies
I think the editor: tag is for actual edits, not for AI generated images. I agree that a similar convention would not make sense for AI generated images. Following the editor: convention for AI generated images could easily lead to cases where people misattribute contributions, which would be bad for both the original artist and the one using AI.
I’m @Background Pony #ECE7, just on a different computer.
Posted Report
Background Pony #38F5
@Cocaine
It is not even slightly embarrassing that visually, conceptually, and technologically interesting creations that humans have decided to share with one another, depicting subjects relevant to the topic of the site, are accepted.
Of course, for the ones that really are “crap”, it’s a little embarrassing. But that’s true for non-AI submissions, too.
Posted Report
Cocaine
Non-Fungible Trixie -

Princess Coke
@Background Pony #38F5
Every single ai generated pic is on the same exact level of shittiness as nft monkeys and art tracers. There is zero difference.
And it happens that nft bros are the ones who really love to promote ai crap. Like was majority of them. U know why? Because it’s the same shit. It’s an attempt to turn art into pure commodity devoid of any meaning.
Posted Report
Edited
Interested in advertising on Derpibooru? Click here for information!
Champions of Equestria

Help fund the $15 daily operational cost of Derpibooru - support us financially!

Edit Preview
Syntax quick reference: **bold** *italic* ||hide text|| `code` __underline__ ~~strike~~ ^sup^ %sub%

Detailed syntax guide