AI Art: It is criminal to not post your prompt

The Something Awful Forums > Main > General Bullshit > AI Art: It is criminal to not post your prompt

«‹›436 »

Monglo: Mar 19, 2015

This must be how the Butlerian Jihad started.

# ? Feb 23, 2023 06:39

Adbot: ADBOT LOVES YOU

# ? Jun 5, 2024 10:51

Ruffian Price: Sep 17, 2016

KakerMix posted:

He can't stop submissions from coming in that have been AI-assisted and he's going to approve ones that have been.

Yeah, all he can really do about this is pretend he can tell immediately and post a hardline stance on his website in the hope of deterring at least some of the spam. He should try that approach

# ? Feb 23, 2023 07:52

Duck and Cover: Apr 6, 2007

Doctor Zero posted:

You have to understand what publishing is like, especially in science fiction, fantasy, and horror markets. They get tons of submissions every day. Out of all that maybe 10% is worthy of consideration. That�s not a snobbery thing that�s 90% that aren�t in the right format, or aren�t science fiction, or not containing full sentences etc. Yes, it is that bad. I�ve heard it from many editors and pros.

Now you have every moron that watches the YouTube videos for USE AI TO BECOME RICH AND FAMOUS WITH NO EFFORT and decides that they can be the next Stephen King by using ChatGPT. And it becomes an avalanche of crap. The submitters are essentially DDOSing the magazines.

LOL if you think the editor of an award winning sci-fi magazine is anti-AI. The point isn�t that AI stories are being submitted the point is that they are getting SO much of it that they can�t keep the submissions open.

They are taking a hard stance because otherwise you�d get people going �well, the AI is just my ghost writer� or whatever lame excuse to try to skirt around the rule.

The the part I never considered is that the *humans* would be the problem with no-effort get rich quick scams.

I�m sure his �secret� method is just reading the manuscript. Right now AI can�t write a good story and if you�ve played around with chat GPT it�s pretty apparent that it�s not very good at being creative.

I understand the concerns you have raised about the use of AI-generated content in the publishing industry. It is true that there is a massive influx of submissions in the science fiction, fantasy, and horror markets, making it challenging for publishers to sift through all of them and find worthy stories. However, AI-generated content should not be dismissed outright, as it is still a relatively new technology that is constantly improving.

While it is true that AI-generated stories are not yet on par with the creativity and complexity of human-authored stories, AI-generated content can still be a useful tool for authors and publishers. AI can assist with tasks such as generating story prompts or helping with language and grammar checks, which can save authors time and effort.

Furthermore, it is essential to keep an open mind about the use of AI in publishing, as it can provide opportunities for new voices to enter the industry, and it can help publishers discover unique and innovative stories that they may not have encountered otherwise.

That being said, it is important to recognize the limitations of AI-generated content and to acknowledge that it cannot replace the creativity and nuance that comes from human authorship. Ultimately, the decision to publish a story should rest on its quality and merit, regardless of whether it was written by a human or generated by AI.

# ? Feb 23, 2023 08:05

Prolonged Panorama: Dec 21, 2007; Holy hookrat Sally smoking crack in the alley!

KakerMix posted:

That isn't what he's saying though. Nothing is going to change, taking a ~hard~ stance on it doesn't matter because he can't actually tell. He can't stop submissions from coming in that have been AI-assisted and he's going to approve ones that have been. He doesn't know and it makes people like me go "lol you're just like everyone on twitter" and lets me safely stereotype him along with the rest and not take him seriously.

I'll repost a quote from the Book Barn Culture thread. The Culture is a series of science fiction novels by the late Iain M Banks, and the setting includes hyper-smart benevolent AIs, called "Minds," who interact with the more typically human/android/robot protagonists.

The passage below is a conversation between an AI Mind (via a corporeal humanoid avatar), and a famous human (well, fuzzy centaur thing, but you know) composer. The Mind has admitted that it could write a symphony in the composer's style that would fool any critic in to thinking it had been by the human, and that it could do so very quickly.

Iain M Banks, in Look to Windward posted:

�So what," the Chelgrian asked, "is the point of me or anybody else writing a symphony, or anything else?"

The avatar raised its brows in surprise. "Well, for one thing, if you do it, it's you who gets the feeling of achievement."

"Ignoring the subjective. What would be the point for those listening to it?"

"They'd know it was one of their own species, not a Mind, who created it."

"Ignoring that, too; suppose they weren't told it was by an AI, or didn't care."

"If they hadn't been told then the comparison isn't complete; information is being concealed. If they don't care, then they're unlike any group of humans I've ever encountered."

"But if you can�"

"Ziller, are you concerned that Minds�AIs, if you like�can create, or even just appear to create, original works of art?"

"Frankly, when they're the sort of original works of art that I create, yes."

"Ziller, it doesn't matter. You have to think like a mountain climber."

"Oh, do I?"

"Yes. Some people take days, sweat buckets, endure pain and cold and risk injury and�in some cases�permanent death to achieve the summit of a mountain only to discover there a party of their peers freshly arrived by aircraft and enjoying a light picnic."

"If I was one of those climbers I'd be pretty damned annoyed."

"Well, it is considered rather impolite to land an aircraft on a summit which people are at that moment struggling up to the hard way, but it can and does happen. Good manners indicate that the picnic ought to be shared and that those who arrived by aircraft express awe and respect for the accomplishment of the climbers.

"The point, of course, is that the people who spent days and sweated buckets could also have taken an aircraft to the summit if all they'd wanted was to absorb the view. It is the struggle that they crave. The sense of achievement is produced by the route to and from the peak, not by the peak itself. It is just the fold between the pages." The avatar hesitated. It put its head a little to one side and narrowed its eyes. "How far do I have to take this analogy, Cr. Ziller?�

Submitting AI written or assisted pieces to a publication that has asked you not to is just... rude. The same way submitting a photograph that's been photoshop-filtered to a painting contest is rude. Or a 3D print of a digital sculpt to a traditional sculpture showcase. Maybe nobody will be able to tell, but that's not the point. You're violating the (nebulous, subjective) spirit of the thing. It's a different class of activity, a different form of art. One is not lesser or greater than the other, they're just different processes, even if the result can look the same.

Clarkesworld magazine is a human construct, a relationship between readers, editors, and contributors. They decide what they want to create, publish, and read. Rude people won't be arrested for spamming works that the community has decided it has no interest in, but they will be shamed, shunned, etc, if and when their dishonesty is uncovered. People who want to create, publish, and read AI material can make their own publications. Just like every other niche hobby, market, scene, etc.

Prolonged Panorama fucked around with this message at 08:56 on Feb 23, 2023

# ? Feb 23, 2023 08:20

KakerMix: Apr 8, 2004; 8.2 M.P.G.

Prolonged Panorama posted:

~Real Life~ isn't a novel. I'm being pragmatic and realistic.

If he wants to declare he can somehow tell AI has "been involved at all in the process. Assisted, written, co-written... whatever you want to call it" and accept or decline submissions on that then he is free to. Reality doesn't work that way. The old ways are dead and we all have to adjust.

People are rude, unfortunately though shaming them doesn't seem to work very well. Ever.

# ? Feb 23, 2023 08:58

Tunicate: May 15, 2012

funny seeing a twitter thread trying to do a ~deep psychoanalysis~ of motivation when the answer is "people submit stories to magazines and anthologies in order to get paid".

most famous example in SF/F is Zelazny, who was an excellent writer and also incredibly mercenary when it came to money. The joke about him was 'what does Zelazny write? Whatever gets him fifteen cents a word'.

Anyway, one day two separate editors asked him to contribute a story to their themed anthologies - one was stories set in bars, and one was about unicorns. He mentioned it to his friend GRRM, who said 'yeah and I just got a request for a story about chess, lol you might as well write a story about a unicorn playing chess in a bar to get paid three times'.

In response, Zelazny wrote a story about a unicorn playing chess in a bar, got paid three times, and won a Hugo.

# ? Feb 23, 2023 09:38

Aertuun: Dec 18, 2012

Aertuun posted:

I have a Virus question for the thread which I can't really find answered anywhere else on the Internet. Hopefully someone here has already figured all this out.

In summary, had Windows Defender flag up some warnings on Stable Diffusion. I am/was using the AUTOMATIC1111 webui.

Windows Defender found a trojan (Trojan:Win32/Sirefef!cfg) in each of the eight ControlNet models that I had downloaded from: https://huggingface.co/lllyasviel/ControlNet/tree/main/models

Objective Action posted:

I will note the one upside to keeping the full models is they work on any base 1.5 model I've tried. The stripped down ones were occasionally doing weird poo poo with some models. YMMV of course.

It might be true that the larger models are better in some ways, however given the source version of them consistently flags as being infected by a trojan, is it sensible to use them? Unless I'm missing something.

# ? Feb 23, 2023 09:59

porfiria: Dec 10, 2008; by Modern Video Games

KakerMix posted:

That isn't what he's saying though. Nothing is going to change, taking a ~hard~ stance on it doesn't matter because he can't actually tell. He can't stop submissions from coming in that have been AI-assisted and he's going to approve ones that have been. He doesn't know and it makes people like me go "lol you're just like everyone on twitter" and lets me safely stereotype him along with the rest and not take him seriously.

I mean, depending on your definition of AI-assisted he can probably tell. But say he can't.

Clarkesworld is a quite well known sci fi magazine, but we're talking a niche part of a niche industry here. Maybe I'm completely off but I doubt they get more than a few thousand submissions a month, which means that 700 (!) fakes in February is a huge increase in volume. If you read between the lines here, the real issue is that there's simply no way to actually wade through all of the stuff. If you're right and he can't tell them apart, there's no way for a magazine like Clarkesworld to accept open submissions, which means a niche part of a niche industry becomes even more insular, which sucks. I don't see any upside to this, personally.

Edit:

Ah, I was right:
https://techcrunch.com/2023/02/21/c...t%20plagiarism.
They normally get about 1,000 submissions a month. So how is he able to tell there were 700 AI stories in February? Well, probably because he got a massive and unprecedented doubling of incoming stories in a single month!

porfiria fucked around with this message at 10:08 on Feb 23, 2023

# ? Feb 23, 2023 10:00

Ruffian Price: Sep 17, 2016

The worst part of this signal boost is that now outside of spam written for a quick buck he'll also start getting techbro spam written out of spite, just to annoy and waste time

# ? Feb 23, 2023 10:24

busalover: Sep 12, 2020

He could use AI to detect most of the obvious stuff, and go from there.

# ? Feb 23, 2023 10:33

reignonyourparade: Nov 15, 2012

porfiria posted:

I mean, depending on your definition of AI-assisted he can probably tell. But say he can't

KakerMix's definition or your definition or my definition doesn't matter, what matters is HIS definition, which he has given us. His definition is any ai involved anywhere in the process.

So no as a matter of fact, he cannot tell the difference between a story where the prompt came from an AI vs came from one of the many many non-ai random prompt generators vs came from a reddit story post vs the author had the idea on his own because of course he can't. He can't tell the difference between a story outlined by an ai and one entirely written by a human. He probably can't tell the difference between one proofread by an AI and one proofread by a human but in that one case eh maybe he can sometimes catch some of them (though probably with plenty of false positives.)

"He knows he got a bunch of ai written pieces" was never at any point in doubt. What was in doubt was his ability to police the maximalist standard he has set.

# ? Feb 23, 2023 10:40

Doctor Zero: Sep 21, 2002; Would you like a jelly baby?
It's been in my pocket through 4 regenerations,
but it's still good.

reignonyourparade posted:

KakerMix's definition or your definition or my definition doesn't matter, what matters is HIS definition, which he has given us. His definition is any ai involved anywhere in the process.

So no as a matter of fact, he cannot tell the difference between a story where the prompt came from an AI vs came from one of the many many non-ai random prompt generators vs came from a reddit story post vs the author had the idea on his own because of course he can't. He can't tell the difference between a story outlined by an ai and one entirely written by a human. He probably can't tell the difference between one proofread by an AI and one proofread by a human but in that one case eh maybe he can sometimes catch some of them (though probably with plenty of false positives.)

"He knows he got a bunch of ai written pieces" was never at any point in doubt. What was in doubt was his ability to police the maximalist standard he has set.

The policy he set doesn't have anything to do with an ability to "police" submissions, but rather the need to set boundaries for submissions. Just like the submission guidelines state he doesn't want Star Trek-like story resolutions that are hand waving or technobabble, the stance on "no AI used" in the writing process is to inform the human beings submitting the work what is acceptable. Of course he can't tell if the idea from a story came from an AI any more than he'd be able to know if an idea came from a dice roll against a chart. He can't tell if a story was edited by AI any more than he could tell if someone paid a copy editor to proof the story. Again, what he's doing is making clear what kinds of stories he will accept and which kinds he won't.

The whole blog post talks about how human beings ignore all that and submit anyway, so he already knows that stating he isn't accepting AI stories isn't going to stop the get rich quick idiots from submitting. However he still has to state what the rules are in his submission guidelines. It's just keeping the honest honest, and he knows that.

That statement in the guidelines doesn't make him anti-AI any more than stating he doesn't want Star Trek-ish stories makes him anti-Star Trek. He doesn't want historical romance stories either, that doesn't make him anti-historical romance either.

Unlike people who actually are anti-AI fear mongers, he doesn't call for AI to be banned. He is banning people who egregiously ignore the submission guidelines, because that's what he's always done, not solely because they are submitting AI stories.

Again, as to his detection method, he admits the detection tools are useless. I'm 99% positive his method is to simply read the story. He reads thousands of stories, so I'm sure he can tell an AI story from a human story easily, just like I knew that Duck and Cover's reply was AI generated before I even finished the first sentence. But making sure takes time to read. Time that is taken away from people who are actually following the submission guidelines. THIS is the whole point of his article. He doesn't go into detail about how he can tell not because it's some arcane and special method, but because by stating what gives AI away will just give a list of things to avoid to the people submitting AI stories.

Doctor Zero fucked around with this message at 13:51 on Feb 23, 2023

# ? Feb 23, 2023 13:47

busalover: Sep 12, 2020

Sketchfab's terms of use are about to change

quote:

We are writing to let you know that we will be introducing some changes to our Terms of Use next month.

If you wish to protect your uploads from usage by generative AI programs, you may tag your models �NoAI�.

You (and other users) agree not to use uploads marked as �NoAI� in datasets for, in the development of, or as inputs to generative AI programs.

If you make your work available under a license that permits redistribution (including as part of a larger work), for example one of the Creative Commons licenses or the standard Sketchfab License Agreement, you agree to tag any of your models that have been made with the use of generative AI programs as �CreatedWithAI�.

Sketchfab agrees not to use any of your uploads in datasets for, in the development of, or as inputs to generative AI programs.

Sketchfab agrees not to license any of your uploads to third parties for use in datasets for, in the development of, or as inputs to generative AI programs.

Terms of Use changes will go into effect on March 23, 2023. You can see the full text for the upcoming changes to Sketchfab�s Terms of Use here.

# ? Feb 23, 2023 13:59

Aertuun: Dec 18, 2012

The default installs of Codeformer in both InvokeAI and the Automatic1111 webui get flagged for trojans by Windows Defender, but only after you unarchive the .pth file.

Trojan:Win32/Sirefef!cfg
\InvokeAI\models\codeformer\archive\data\94338023080784
\InvokeAI\models\codeformer\archive\data\94338067080304

# ? Feb 23, 2023 15:40

Objective Action: Jun 10, 2007

Aertuun posted:

The default installs of Codeformer in both InvokeAI and the Automatic1111 webui get flagged for trojans by Windows Defender, but only after you unarchive the .pth file.

Trojan:Win32/Sirefef!cfg
\InvokeAI\models\codeformer\archive\data\94338023080784
\InvokeAI\models\codeformer\archive\data\94338067080304

Interesting and concerning. I pulled the block and ran them through VirusTotal and Microsoft was the only one that flagged it.

While it seems a likely false positive I've converted my models over to .safetensor versions just because there isn't a good reason not to be safer and now Microsoft comes back clean as well.

# ? Feb 23, 2023 16:01

LASER BEAM DREAM: Nov 3, 2005; Oh, what? So now I suppose you're just going to sit there and pout?

If anyone else is also just getting started, or looking to grow their skills, I have really enjoyed this youtube channel

https://www.youtube.com/@OlivioSarikas

Olivio is knowledgeable, concise, and bleeding edge. I used his channel for my entire Automatic1111 install and he's already made 3 increasingly advanced Control Net tutorials.

# ? Feb 23, 2023 17:00

Doctor Zero: Sep 21, 2002; Would you like a jelly baby?
It's been in my pocket through 4 regenerations,
but it's still good.

LASER BEAM DREAM posted:

If anyone else is also just getting started, or looking to grow their skills, I have really enjoyed this youtube channel

https://www.youtube.com/@OlivioSarikas

Olivio is knowledgeable, concise, and bleeding edge. I used his channel for my entire Automatic1111 install and he's already made 3 increasingly advanced Control Net tutorials.

Seconding. He's really good. He explains everything well, doesn't assume people know things, and doesn't have a negative attitude.

# ? Feb 23, 2023 17:04

Lucid Dream: Feb 4, 2003; That boy ain't right.

https://clips.twitch.tv/SpikyMotionlessTurnipUWot-b-HWCLMoAk480xr_

# ? Feb 23, 2023 17:48

slurm: Jul 28, 2022; by Hand Knit

Duck and Cover posted:

I understand the concerns you have raised about the use of AI-generated content in the publishing industry. It is true that there is a massive influx of submissions in the science fiction, fantasy, and horror markets, making it challenging for publishers to sift through all of them and find worthy stories. However, AI-generated content should not be dismissed outright, as it is still a relatively new technology that is constantly improving.

While it is true that AI-generated stories are not yet on par with the creativity and complexity of human-authored stories, AI-generated content can still be a useful tool for authors and publishers. AI can assist with tasks such as generating story prompts or helping with language and grammar checks, which can save authors time and effort.

Furthermore, it is essential to keep an open mind about the use of AI in publishing, as it can provide opportunities for new voices to enter the industry, and it can help publishers discover unique and innovative stories that they may not have encountered otherwise.

That being said, it is important to recognize the limitations of AI-generated content and to acknowledge that it cannot replace the creativity and nuance that comes from human authorship. Ultimately, the decision to publish a story should rest on its quality and merit, regardless of whether it was written by a human or generated by AI.

Lol it doesn't yet have the heart of the poster as seen here

# ? Feb 23, 2023 17:52

Mozi: Apr 4, 2004; Forms change so fast
Time is moving past
Memory is smoke
Gonna get wider when I die; Nap Ghost

Challenge: Have ChatGPT write a poem that contains no rhymes.

# ? Feb 23, 2023 19:33

LASER BEAM DREAM: Nov 3, 2005; Oh, what? So now I suppose you're just going to sit there and pout?

Someone made an OpenPose extension for Automatic1111!

Check under Extensions to install it. It's pretty basic, but its better than my sketchwork!

# ? Feb 23, 2023 19:46

StarkRavingMad: Sep 27, 2001; Yams Fan

Mozi posted:

Challenge: Have ChatGPT write a poem that contains no rhymes.

Harder than I thought! I tried telling it to write a poem with no rhymes. Then I yelled at it for rhyming, and it apologized and gave me another rhyming poem. I tried other stuff like telling it to write in the style of certain poets, telling it to give me another verse for "When I Heard the Learn�d Astronomer", etc. All rhymes.

Finally figured out the trick, you have to avoid the term "poem," it thinks all poems rhyme. Free verse or blank verse work:

# ? Feb 23, 2023 20:10

Mozi: Apr 4, 2004; Forms change so fast
Time is moving past
Memory is smoke
Gonna get wider when I die; Nap Ghost

Ah, that's the trick. Yeah I was just banging my head against 'poem=rhymes' and getting frustrated. I wonder what needs to happen technologically for it to be able to take feedback like "You used 'tree' and 'free' as rhymes, I asked for no rhymes" and change what it gives you instead of just going 'I apologize, let me try that again <more rhymes>'

# ? Feb 23, 2023 20:12

cYn: Apr 1, 2008

Ihmemies posted:

So I'm making a city builder, and of course it must have advisors. Advisor system is a very important system, especially it must be done before you can even place roads in the game

Concept art style. Broad strokes. Head and shoulders portrait. Female doctor. Homely. Wearing doctor's clothing. Eyeglasses. Light background. --quality .5 --ar 2:3 --style 4c

x many times. Well. Better than the nothing I can myself draw! hard to get it to use a cohesive style, it varies from image to image. Sometimes it wants to do more abstract, sometimes less abstract!

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

# ? Feb 23, 2023 20:55

Cousin Todd: Jul 3, 2007; Grimey Drawer

cYn posted:

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

(USER WAS PUT ON PROBATION FOR THIS POST)

# ? Feb 23, 2023 20:58

Boba Pearl: Dec 27, 2019; by Athanatos

Put me down as the artist, I drew it all pro bono.

# ? Feb 23, 2023 21:06

Boba Pearl: Dec 27, 2019; by Athanatos

There's literally 0 way to prove that art was done by AI, so worst case scenario, credit yourself as the artist, but also there's absolutely no copyright lawsuits that'll hit your game because you released games commercially with AI art, because it's already happened in both anime, netflix, and video games.

# ? Feb 23, 2023 21:07

Megazver: Jan 13, 2006

cYn posted:

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

https://arstechnica.com/information-technology/2023/02/us-copyright-office-withdraws-copyright-for-ai-generated-comic-artwork/

The art pieces itself might not get copyright, but everything else still will. Do you really care if someone can reuse the portraits?

# ? Feb 23, 2023 21:13

KwegiboHB: Feb 2, 2004; nonconformist art brut
Negative prompt: amenable, compliant, docile, law-abiding, lawful, legal, legitimate, obedient, orderly, submissive, tractable
Steps: 32, Sampler: DPM++ 2M Karras, CFG scale: 11, Seed: 520244594, Size: 512x512, Model hash: 99fd5c4b6f, Model: seekArtMEGA_mega20

cYn posted:

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

Edit: The above article is news to me, let me read it.

No.
I've been following this closely, that graphic novel you are talking about is called Zarya of the Dawn and their copyright remains fully intact. There is a purposeful campaign to poison the well around AI Art, including at least one PR firm that I've tracked down. The reason it feels like people are screaming nonsense without bothering to listen to another side is that they're getting paid to do just that, on purpose.

Pay yourself a dollar for the art you created for yourself. Artist commissioned.

Boba Pearl, I finally got my webui reset and working again and can spend the rest of the day on this Block Merge project. I have a few grids started for comparison but so far it's just "archer style landscape portrait of city at night" vs "landscape portrait of city at night" to test the effects of the keyword on line strength. Do you have anything particular you'd like for the test grids to compare? Give me a prompt and I'll run sets using it so you can see the difference.

KwegiboHB fucked around with this message at 21:30 on Feb 23, 2023

# ? Feb 23, 2023 21:23

Boba Pearl: Dec 27, 2019; by Athanatos

Here's a deceptively hard one when using img2img

The prompt is "A teal open door over aqua concrete steps, with a green brick wall, and a blue door frame, archer style, nomura tetsuya <lora:nomura tetsuya:0.6>, flat bright colors."
The negative prompt is "beeple, vray render, 3D, Daz3d, c4d, 3ds, bad, ugly, photorealistic, iray, octane"
I've done a few, and eventually gave up, this image is deceptively difficult for some reason:

# ? Feb 23, 2023 21:55

Cousin Todd: Jul 3, 2007; Grimey Drawer

I'm not surprised, that reference image is just a bunch of garbage information visually, it doesn't really look like the right dimensions or angles for a door.

# ? Feb 23, 2023 21:59

Boba Pearl: Dec 27, 2019; by Athanatos

It's a close up of the bottom half of a door, but that makes sense, that's the problem a lot of time with this method, but the good thing is, is it's really good at showing early when a scene might be flawed.

# ? Feb 23, 2023 22:16

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

cYn posted:

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

You can indeed copyright a work partially created with AI. The AI images themselves are not copyrightable, just like public domain photos of national parks aren't copyrighted, but if you created a music video with those photos in the background you would own the video.
https://twitter.com/icreatelife/status/1628454855315521536?s=20

Here is the IRS's most recent letter to the author, dated Feb 21st:

quote:

Dear Mr. Lindberg:

The United States Copyright Office has reviewed your letter dated November 21, 2022,
responding to our letter to your client, Kristina Kashtanova, seeking additional information
concerning the authorship of her work titled Zarya of the Dawn (the �Work�). Ms. Kashtanova
had previously applied for and obtained a copyright registration for the Work, Registration
# VAu001480196. We appreciate the information provided in your letter, including your
description of the operation of the Midjourney�s artificial intelligence (�AI�) technology and
how it was used by your client to create the Work.

The Office has completed its review of the Work�s original registration application and
deposit copy, as well as the relevant correspondence in the administrative record.1 We conclude
that Ms. Kashtanova is the author of the Work�s text as well as the selection, coordination, and
arrangement of the Work�s written and visual elements. That authorship is protected by
copyright. However, as discussed below, the images in the Work that were generated by the
Midjourney technology are not the product of human authorship. Because the current
registration for the Work does not disclaim its Midjourney-generated content, we intend to
cancel the original certificate issued to Ms. Kashtanova and issue a new one covering only the
expressive material that she created.

[...]

Reuters Article (with a kinda misleading title)

quote:

Feb 22 (Reuters) - Images in a graphic novel that were created using the artificial-intelligence system Midjourney should not have been granted copyright protection, the U.S. Copyright Office said in a letter seen by Reuters.

"Zarya of the Dawn" author Kris Kashtanova is entitled to a copyright for the parts of the book Kashtanova wrote and arranged, but not for the images produced by Midjourney, the office said in its letter, dated Tuesday.

The decision is one of the first by a U.S. court or agency on the scope of copyright protection for works created with AI, and comes amid the meteoric rise of generative AI software like Midjourney, Dall-E and ChatGPT.

e: whoops I missed megazver's post about the same thing

# ? Feb 23, 2023 22:20

AARD VARKMAN: May 17, 1993

# ? Feb 23, 2023 22:35

Lucid Dream: Feb 4, 2003; That boy ain't right.

AARD VARKMAN posted:

I'm so happy for them.

# ? Feb 23, 2023 22:37

Play: Apr 25, 2006; Strong stroll for a mangy stray

StarkRavingMad posted:

Harder than I thought! I tried telling it to write a poem with no rhymes. Then I yelled at it for rhyming, and it apologized and gave me another rhyming poem. I tried other stuff like telling it to write in the style of certain poets, telling it to give me another verse for "When I Heard the Learn’d Astronomer", etc. All rhymes.

Finally figured out the trick, you have to avoid the term "poem," it thinks all poems rhyme. Free verse or blank verse work:

Well I think the poets are safe for the moment. That is some straight up doggerel

# ? Feb 23, 2023 22:38

cYn: Apr 1, 2008

RPATDO_LAMD posted:

You can indeed copyright a work partially created with AI. The AI images themselves are not copyrightable, just like public domain photos of national parks aren't copyrighted, but if you created a music video with those photos in the background you would own the video.
https://twitter.com/icreatelife/status/1628454855315521536?s=20

Here is the IRS's most recent letter to the author, dated Feb 21st:

Reuters Article (with a kinda misleading title)

e: whoops I missed megazver's post about the same thing

Alls I'm saying is that it's still in flux and it could go either way. I wouldn't rely on it being viable in the long term just yet.
Don't assume I'm opposed to AI image generation, I just know that there are a lot of details yet to be figured out regarding copyright, licencing and consent.

My job entails keeping the studio I work for, up to date on new tech and obviously AI is a big part of that. That's why I'm in this thread, I'm figuring stuff out and making stuff and exploring the potential. But I wouldn't suggest we sell anything generated by it just yet.

# ? Feb 23, 2023 22:40

Ihmemies: Oct 6, 2012

cYn posted:

You're not planning on releasing it commercially are you? Currently anything generated by AI has potential to be crushed by copyright laws.
The precedent has already been set regarding graphic novels using image synthesis.
Just comission an artist.

I'll think about it if/when I ever get a functional game which is fun to play, and works, and even slightly resembles the original ideas I had. Meanwhile I'm going to use some placehoder graphics while developing the game :v:

I don't know how to comission an artis without a budget as an Computer Science student. Sounds like something you'd need money for. Lots of money.

Anyways, how do weights work in midjourney? I wanted a 2d cad drawing of an oil pan. Well it produced a 3d drawing.

How about giving more weight to 2d? "2-dimensional::10 cad drawing of an combustion engine oil pan." Bitch forgets all about the cad and oil pan and produces this:

# ? Feb 23, 2023 22:41

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

Boba Pearl posted:

Here's a deceptively hard one when using img2img

The prompt is "A teal open door over aqua concrete steps, with a green brick wall, and a blue door frame, archer style, nomura tetsuya <lora:nomura tetsuya:0.6>, flat bright colors."
The negative prompt is "beeple, vray render, 3D, Daz3d, c4d, 3ds, bad, ugly, photorealistic, iray, octane"
I've done a few, and eventually gave up, this image is deceptively difficult for some reason:

[snip]

you should try out ControlNet, it's really great for stuff like this where you want the generated image to closely follow some input.
I used the canny edge detection module and got this:

edge detection:

with controlnet + img2img:

with just ControlNet in txt2img mode:

both of these using your same prompt minus the LoRa, on sd1.5
you can see it's a lot better at sticking to the same composition as the input image (but it still doesn't know where to put a door)

that was at 95% guidance strength, you can turn it up and down to get the model to follow the lines more or less.

# ? Feb 23, 2023 22:44

Adbot: ADBOT LOVES YOU

# ? Jun 5, 2024 10:51

KakerMix: Apr 8, 2004; 8.2 M.P.G.

cYn posted:

Alls I'm saying is that it's still in flux and it could go either way. I wouldn't rely on it being viable in the long term just yet.
Don't assume I'm opposed to AI image generation, I just know that there are a lot of details yet to be figured out regarding copyright, licencing and consent.

My job entails keeping the studio I work for, up to date on new tech and obviously AI is a big part of that. That's why I'm in this thread, I'm figuring stuff out and making stuff and exploring the potential. But I wouldn't suggest we sell anything generated by it just yet.

"What? I made this"

# ? Feb 23, 2023 22:44

The Something Awful Forums > Main > General Bullshit > AI Art: It is criminal to not post your prompt

«‹›436 »