AI Generated Imagery

Everything else!

Moderators: Bakhtosh, EvilHomer3k

User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

MidJourney is surprisingly bad at two people.
I have but one advocate.
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

AI-wielding artist wins first place at Colorado State Fair
A synthetic media artist named Jason Allen entered AI-generated artwork into the Colorado State Fair fine arts competition and announced last week that he won first place in the Digital Arts/Digitally Manipulated Photography category, Vice reported Wednesday based on a viral tweet.

Allen used Midjourney—a commercial image synthesis model available through a Discord server—to create a series of three images. He then upscaled them, printed them on canvas, and submitted them to the competition in early August. To his delight, one of the images (titled Théåtre D'opéra Spatial) captured the top prize, and he posted about his victory on the Midjourney Discord server on Friday.

Allen's victory prompted lively discussions on Twitter, Reddit, and the Midjourney Discord server about the nature of art and what it means to be an artist. Some commenters think human artistry is doomed thanks to AI and that all artists are destined to be replaced by machines. Others think art will evolve and adapt with new technologies that come along, citing synthesizers in music. It's a hot debate that Wired covered in July.

There's also the fairness element since it isn't clear if Allen told the judges about his use of image synthesis, though some Twitter users have reportedly contacted the judges and discovered that they didn't know. Curiously, the art was considered good enough to fool human artists, and someone on Twitter joked that it settled the debate over "whether AI art is art."
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

MidJourney is now testing a photorealistic mode that's pretty amazing. Besides making whatever you prompt as normal but without MidJourney's usual art style, you can also name a celebrity and some of the results are good enough to look real. Want Alison Brie as a cute dryad? It nails it. Not sure how that will factor into their usual tools when it's out of testing.
I have but one advocate.
User avatar
Jaymann
Posts: 20567
Joined: Mon Oct 25, 2004 7:13 pm
Location: California

Re: AI Generated Imagery

Post by Jaymann »

Max Peck wrote: Wed Aug 31, 2022 5:25 pm AI-wielding artist wins first place at Colorado State Fair
A synthetic media artist named Jason Allen...
Made me think the artist was synthetic.
Jaymann
]==(:::::::::::::>
Leave no bacon behind.
User avatar
Isgrimnur
Posts: 84864
Joined: Sun Oct 15, 2006 12:29 am
Location: Chookity pok
Contact:

Re: AI Generated Imagery

Post by Isgrimnur »

Enlarge Image
It's almost as if people are the problem.
User avatar
Daehawk
Posts: 65687
Joined: Sat Jan 01, 2005 1:11 am

Re: AI Generated Imagery

Post by Daehawk »

--------------------------------------------
I am Dyslexic of Borg, prepare to have your ass laminated.
I guess Ray Butts has ate his last pancake.
http://steamcommunity.com/id/daehawk
"Has high IQ. Refuses to apply it"
When in doubt, skewer it out...I don't know.
User avatar
Gryndyl
Posts: 4746
Joined: Tue Jan 11, 2005 1:12 pm
Location: Washington

Re: AI Generated Imagery

Post by Gryndyl »

JetFred wrote: Thu Aug 11, 2022 7:03 pm A friend and I cannot stop making D&D locations, items, creatures, and portraits with MidJourney. These selections are typical, not exceptional. I'd estimate about an hour or two of tweaking to get these ten images. Click for full.
Really nice! I've only gotten to play with these a little bit, basically the 20 free uses that midjourney gives. I'm experimenting with using it to generate book covers. I was pretty pleased with what I got with my brief experience (favorite is the image below) but I wasn't managing to get anything like some of those pics you got. Any tips on giving prompts?

Image
User avatar
Unagi
Posts: 28195
Joined: Wed Sep 20, 2006 5:14 pm
Location: Chicago

Re: AI Generated Imagery

Post by Unagi »

Jaymann wrote: Wed Aug 31, 2022 6:14 pm
Max Peck wrote: Wed Aug 31, 2022 5:25 pm AI-wielding artist wins first place at Colorado State Fair
A synthetic media artist named Jason Allen...
Made me think the artist was synthetic.
Smart is understanding the artist wasn’t synthetic.
Wise is understandinge the artist was synthetic.
User avatar
Carpet_pissr
Posts: 20793
Joined: Thu Nov 04, 2004 5:32 pm
Location: Columbia, SC

Re: AI Generated Imagery

Post by Carpet_pissr »

JetFred wrote: Wed Aug 31, 2022 6:04 pm Want Alison Brie as a cute dryad? It nails it. Not sure how that will factor into their usual tools
Setup…joke…overload….too….much….AAAAUGGHH! :

YespleasemichaelscottibetitdoessilenceofthelambsHARDLY…
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Stable Diffusion Brings Local AI Art Generation to Your PC
AI-generated artwork is incredibly popular now. It’s now possible to generate photorealistic images right on your PC, without using external services like Midjourney or DALL-E 2.

Stability AI is a tech startup developing the “Stable Diffusion” AI model, which is a complex algorithm trained on images from the internet. Following a test version available to researchers, the company has officially released the Stable Diffusion model, which can be used to create images from text prompts. Unlike Midjourney and other models/generators, Stable Diffusion aims to create photorealistic images first and foremost — something that has already led to controversy over “deepfake” content. However, it can also be configured to mimic the style of a given artist.

Stable Diffusion is unique because it can run with a typical graphics card, instead of using remote (and expensive) servers to generate images. Stability AI recommends using NVIDIA graphics cards right now, but full support for AMD and Apple Silicon is in the works.
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Gryndyl wrote: Sun Sep 04, 2022 12:08 am ...Any tips on giving prompts?
It helps to look at the community showcase to see how people are using descriptors like "cinematic lighting" or "photography," etc. I'm assuming you've searched for and read the brief manual explaining things like weighting (I don't do that much) and setting the stylize or chaos commands. But to be honest, I tend to just keep making variations and rerolling until I get something closer to what I want. That image is really cool, but it also shows why the community has been wearing out the --testp command for photorealism because it kills the "Midjourney style" of adding noise that works for detail like foliage or clouds but makes a mess of buildings.
I have but one advocate.
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Yeah, Midjourney's new photorealism mode goes straight to deepfake, like it or not. As long as you persist until it gets the eyes right, it's pretty convincing.
I have but one advocate.
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Here ya go. I recast Half-Life.

Enlarge ImageEnlarge Image
Enlarge Image
Enlarge Image
Enlarge Image
Enlarge ImageEnlarge Image
Enlarge Image
I have but one advocate.
User avatar
Unagi
Posts: 28195
Joined: Wed Sep 20, 2006 5:14 pm
Location: Chicago

Re: AI Generated Imagery

Post by Unagi »

OK, that's kinda amazing.

That's all your own 'work' ?

Request: make a Back to the Future Doc that is actually Walter White of Breaking Bad... ?
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Unagi wrote: Mon Sep 05, 2022 6:11 pm OK, that's kinda amazing.

That's all your own 'work' ?

Request: make a Back to the Future Doc that is actually Walter White of Breaking Bad... ?
Yes, thank you, I got bored last night and did those.

Let me see what I can come up with...
I have but one advocate.
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Unagi wrote: Mon Sep 05, 2022 6:11 pm Request: make a Back to the Future Doc that is actually Walter White of Breaking Bad... ?
Alright, just to be clear, this isn't like Artbreeder. I'm not giving it uploaded images to reference. Had a hard time deciding which way to go with this. With time, I could probably get something closer to the composition of a scene from the movie, but here's what I thought might serve as a little spread of the range of things Midjourney can do right now, not counting the more fantasy and stylized stuff I was showing in my D&D-flavored post above.

The freakin' obvious:
Enlarge Image

The more subtle:
Enlarge Image

The straight up deepfake:
Enlarge Image

And one where...I turned the RV into the time machine?
Enlarge Image
I have but one advocate.
User avatar
Unagi
Posts: 28195
Joined: Wed Sep 20, 2006 5:14 pm
Location: Chicago

Re: AI Generated Imagery

Post by Unagi »

Bravo!!!

That's just amazing, honestly - I can see what has artists upset.
User avatar
JetFred
Posts: 4807
Joined: Fri Oct 15, 2004 11:26 am
Location: In an open field west of a white house

Re: AI Generated Imagery

Post by JetFred »

Unagi wrote: Mon Sep 05, 2022 10:00 pm That's just amazing, honestly - I can see what has artists upset.
Seriously. Consider that I just fed it the nonsense "the entire planet's surface goes to absolute zero and the spirits rise from the ether" with a couple of style commands, and on first attempt it spits out this.

Enlarge Image
I have but one advocate.
User avatar
YellowKing
Posts: 31133
Joined: Wed Oct 13, 2004 2:02 pm

Re: AI Generated Imagery

Post by YellowKing »

I signed up for the DALL-E waiting list. A buddy of mine got accepted and has been making some really cool stuff with it.
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

With Stable Diffusion, you may never believe what you see online again
AI image generation is here in a big way. A newly released open source image synthesis model called Stable Diffusion allows anyone with a PC and a decent GPU to conjure up almost any visual reality they can imagine. It can imitate virtually any visual style, and if you feed it a descriptive phrase, the results appear on your screen like magic.

Some artists are delighted by the prospect, others aren't happy about it, and society at large still seems largely unaware of the rapidly evolving tech revolution taking place through communities on Twitter, Discord, and Github. Image synthesis arguably brings implications as big as the invention of the camera—or perhaps the creation of visual art itself. Even our sense of history might be at stake, depending on how things shake out. Either way, Stable Diffusion is leading a new wave of deep learning creative tools that are poised to revolutionize the creation of visual media.
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Smoove_B
Posts: 56116
Joined: Wed Oct 13, 2004 12:58 am
Location: Kaer Morhen

Re: AI Generated Imagery

Post by Smoove_B »

Unagi wrote: Mon Sep 05, 2022 10:00 pm That's just amazing, honestly - I can see what has artists upset.
Related (sort of) but I follow lots of indie RPG developers on DriveThru and one of them sent out an email today saying they just lost a bunch of clients because of the ability to use AI Generated Imagery in various RPG products. Given what I've seen shared here, I can believe it. It's cool but it never occurred to me how it might be impacting artists specifically l until I saw that email today. Crazy, crazy times.
Maybe next year, maybe no go
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

A door closes, a window opens...

Professional AI whisperers have launched a marketplace for DALL-E prompts
In the past few years, art made by programs like Midjourney and OpenAI’s DALL-E has gotten surprisingly compelling. These programs can translate a text prompt into literally (and controversially) award-winning art. As the tools get more sophisticated, those prompts have become a craft in their own right. And as with any other craft, some creators have started putting them up for sale.

PromptBase is at the center of the new trade in prompts for generating specific imagery from image generators, a kind of meta-art market. Launched earlier this summer to both intrigue and criticism, the platform lets “prompt engineers” sell text descriptions that reliably produce a certain art style or subject on a specific AI platform. When you buy the prompt, you get a string of words that you paste into Midjourney, DALL-E, or another system that you’ve got access to. The result (if it’s a good prompt) is a variation on a visual theme like nail art designs, anime pinups, or “futuristic succulents.”
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Jaymann
Posts: 20567
Joined: Mon Oct 25, 2004 7:13 pm
Location: California

Re: AI Generated Imagery

Post by Jaymann »

Or you can go on discord and see what prompts work well.
Jaymann
]==(:::::::::::::>
Leave no bacon behind.
User avatar
hitbyambulance
Posts: 10624
Joined: Wed Oct 13, 2004 3:51 am
Location: Map Ref 47.6°N 122.35°W
Contact:

Re: AI Generated Imagery

Post by hitbyambulance »

looking at a gallery of generated prompts and their images, - _lots_ of pop-culture mash-ups and tacky bog-standard fantasy/SF tropes. (and almost-naked women - it's a saving grace the gore, porn and white supremacist filters are active by default). it's still a novelty, but i wonder how long til people get really tired of this no-effort 'style'.
User avatar
Daehawk
Posts: 65687
Joined: Sat Jan 01, 2005 1:11 am

Re: AI Generated Imagery

Post by Daehawk »

Dont Fear the Reaper but all images in the video are AI generated.

--------------------------------------------
I am Dyslexic of Borg, prepare to have your ass laminated.
I guess Ray Butts has ate his last pancake.
http://steamcommunity.com/id/daehawk
"Has high IQ. Refuses to apply it"
When in doubt, skewer it out...I don't know.
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

How to Run Stable Diffusion Locally With a GUI on Windows
You can install Stable Diffusion locally on your PC, but the typical process involves a lot of work with the command line to install and use. Fortunately for us, the Stable Diffusion community has solved that problem. Here’s how to install a version of Stable Diffusion that runs locally with a graphical user interface!
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

DALL·E Now Available Without Waitlist
Starting today, we are removing the waitlist for the DALL·E beta so users can sign up and start using it immediately. More than 1.5M users are now actively creating over 2M images a day with DALL·E—from artists and creative directors to authors and architects—with over 100K users sharing their creations and feedback in our Discord community.

Responsibly scaling a system as powerful and complex as DALL·E—while learning about all the creative ways it can be used and misused—has required an iterative deployment approach.

Since we first previewed the DALL·E research to users in April, users have helped us discover new uses for DALL·E as a powerful creative tool. Artists, in particular, have provided important input on DALL·E’s features.

Their feedback inspired us to build features like Outpainting, which lets users continue an image beyond its original borders and create bigger images of any size, and collections—so users can create in all new ways and expedite their creative processes.

Learning from real-world use has allowed us to improve our safety systems, making wider availability possible today. In the past months, we’ve made our filters more robust at rejecting attempts to generate sexual, violent and other content that violates our content policy and built new detection and response techniques to stop misuse.

We are currently testing a DALL·E API with several customers and are excited to soon offer it more broadly to developers and businesses so they can build apps on this powerful system.

We can’t wait to see what users from around the world create with DALL·E. Sign up today and start creating.
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Sudy
Posts: 8412
Joined: Sun Nov 21, 2004 3:11 am
Location: Ontario, Canada

Re: AI Generated Imagery

Post by Sudy »

A few favourites on DALL-E:

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image

I saw a commercial on late night TV. It said, "Forget everything you know about slipcovers." So I did. And it was a load off my mind. Then the commercial tried to sell me slipcovers, and I didn't know what the hell they were. -- Mitch Hedberg
User avatar
Sudy
Posts: 8412
Joined: Sun Nov 21, 2004 3:11 am
Location: Ontario, Canada

Re: AI Generated Imagery

Post by Sudy »

Enlarge Image
Enlarge Image
Enlarge Image
Enlarge Image
Enlarge Image
Enlarge Image
Enlarge Image

I saw a commercial on late night TV. It said, "Forget everything you know about slipcovers." So I did. And it was a load off my mind. Then the commercial tried to sell me slipcovers, and I didn't know what the hell they were. -- Mitch Hedberg
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

AI art generated by the malware behind the cyberattacks on Ukraine is necessarily grim and very wormy
The time of malware generated AI art is upon us, with the first images I've spotted rivalling that of the great HR Geiger in both scale and complexity. Not only has the Twitter user who brought about this unholy union of AI malware art, Greg Linares, made some impressive snapshots of the questionable latent space in image form, they've also used their skills to design a malware generated, cyberpunk themed music album—and it actually slaps.

Though just as the throngs of past wars have inspired great artists to create their masterpieces, so too do the malware artists of our time draw from some dark inspiration.

Linares goes by Laughing_Mantis on Twitter, pulling together their malware-based AI art designs under a dedicated Malware Art profile. Each piece curated here, Linares notes that the art is generated on a heavily modified local version of Stable Diffusion v1.4 and a separate 2.0 install. Basically, he uses "strings and text from inside malware, as well as filenames, and other usable meta data in order to drive the AI art prompts."

It was Linares' glorious collection of images generated using malware from the 'Sandworm group' that caught my eye in the early hours of this morning, the debugging information and test folders for which, Linares explains, were littered with Dune references—hence the Sandworms. These sinuous, cable-draped creations tower over humanity in mist-filled scenes of imminent demise, mouth agape, and hungry for flesh.
https://twitter.com/MalwareArt
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Have you met Loab yet?

🧵: I discovered this woman, who I call Loab, in April. The AI reproduced her more easily than most celebrities. Her presence is persistent, and she haunts every image she touches. CW: Take a seat. This is a true horror story, and veers sharply macabre.
Meet Loab, the AI Art Woman Haunting the Internet
Are there ghosts in our machines? Well, of course not, but a recent viral Twitter thread might have you believing there is something sinister lurking behind your computer screen, just waiting to be unleashed. 

On Sept. 6, the internet was introduced to "Loab," an apparently AI-generated "woman." The internet promptly began calling her "the first cryptid of latent space," "creepy," a "demon" and "a queer icon." There's a lot going on here, so let's explain. 

First, to understand Loab, you need to understand what's happening in AI art.
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image

Enlarge Image
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Smoove_B
Posts: 56116
Joined: Wed Oct 13, 2004 12:58 am
Location: Kaer Morhen

Re: AI Generated Imagery

Post by Smoove_B »

That Loab story is freaking me the f out.
Maybe next year, maybe no go
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Smoove_B wrote: Tue Dec 27, 2022 8:13 pm That Loab story is freaking me the f out.
She found me!

Enlarge Image
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Hrothgar
Posts: 1122
Joined: Wed Oct 13, 2004 11:38 pm
Location: Houston, TX
Contact:

Re: AI Generated Imagery

Post by Hrothgar »

Eww. I went back to the text to pokemon for fun.

Image
Curly Neal from Super Globetrotters

Image
Kratos

Image
Big Boss from MGSV

Image
Die-Hardman from Death Stranding

And, of course,
Image
Paste Pot Pete
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Artists file class-action lawsuit against AI image generator companies
Some artists have begun waging a legal fight against the alleged theft of billions of copyrighted images used to train AI art generators and reproduce unique styles without compensating artists or asking for consent.

A group of artists represented by the Joseph Saveri Law Firm has filed a US federal class-action lawsuit in San Francisco against AI-art companies Stability AI, Midjourney, and DeviantArt for alleged violations of the Digital Millennium Copyright Act, violations of the right of publicity, and unlawful competition.

The artists taking action—Sarah Andersen, Kelly McKernan, Karla Ortiz—"seek to end this blatant and enormous infringement of their rights before their professions are eliminated by a computer program powered entirely by their hard work," according to the official text of the complaint filed to the court.
Alex Champandard, an AI analyst that has advocated for artists' rights without dismissing AI tech outright, criticized the new lawsuit in several threads on Twitter, writing, "I don't trust the lawyers who submitted this complaint, based on content + how it's written. The case could do more harm than good because of this." Still, Champandard thinks that the lawsuit could be damaging to the potential defendants: "Anything the companies say to defend themselves will be used against them."

To Champandard's point, we've noticed that the complaint includes several statements that potentially misrepresent how AI image synthesis technology works. For example, the fourth paragraph of section I says, "When used to produce images from prompts by its users, Stable Diffusion uses the Training Images to produce seemingly new images through a mathematical software process. These 'new' images are based entirely on the Training Images and are derivative works of the particular images Stable Diffusion draws from when assembling a given output. Ultimately, it is merely a complex collage tool."

In another section that attempts to describe how latent diffusion image synthesis works, the plaintiffs incorrectly compare the trained AI model with "having a directory on your computer of billions of JPEG image files," claiming that "a trained diffusion model can produce a copy of any of its Training Images."

During the training process, Stable Diffusion drew from a large library of millions of scraped images. Using this data, its neural network statistically "learned" how certain image styles appear without storing exact copies of the images it has seen. Although in the rare cases of overrepresented images in the dataset (such as the Mona Lisa), a type of "overfitting" can occur that allows Stable Diffusion to spit out a close representation of the original image.

Ultimately, if trained properly, latent diffusion models always generate novel imagery and do not create collages or duplicate existing work—a technical reality that potentially undermines the plaintiffs' argument of copyright infringement, though their arguments about "derivative works" being created by the AI image generators is an open question without a clear legal precedent to our knowledge.
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
User avatar
Unagi
Posts: 28195
Joined: Wed Sep 20, 2006 5:14 pm
Location: Chicago

Re: AI Generated Imagery

Post by Unagi »

So if you train guitar by playing other artists’ music in your basement, you’ll never truly produce a piece of original music.


Obviously that argument needs to be tossed.
User avatar
Max Peck
Posts: 14862
Joined: Fri Aug 05, 2005 8:09 pm
Location: Down the Rabbit-Hole

Re: AI Generated Imagery

Post by Max Peck »

Yeah, it sounds like the people behind the lawsuit don't actually understand how the software works.

As an experiment, I tried generating some images with Stable Diffusion 2.1 using a simple prompt including "In the style of Sarah Andersen." That causes it to generate b&w line art, so I'd say that the prompt is meaningful, but it doesn't really look anything at all like her style, so it would be difficult to say that it is actually mimicking her work. I suppose, with enough detail in a prompt, it might be possible to get something that would make you think it could have been drawn by her, but I believe that would take significant effort.

Enlarge Image

For contrast, to illustrate that "in the style of" is meaningful:

Enlarge Image

Enlarge Image
"What? What? What?" -- The 14th Doctor

It's not enough to be a good player... you also have to play well. -- Siegbert Tarrasch
Post Reply