AI-Replacing-Creatives - Lawsuits & Precedent Setting

[Val the Moofia Boss]

"AI art will put artists out of work!" you have it backwards. We already have a society of bad art and no good artists. The reason why "AI art" is a phenomena now is because most modern artists are awful. Academia spent 100 years destroying art. You need to teach people how to create good art again before you can complain about the machine creating better art. Good artists are very rare and it is very difficult to find them, which is why they are booked months in advance and are expensive to contract. That's why the book industry shifted to using photoshopped book covers. The book industry has accepted that it is okay to not have art. AI art (remixing preexisting graphics) is functionally no different than photoshopped book covers, which we already have now.

The whining about AI art now is superfluous. Most of the paying gigs have either been using it already for composition purposes or were replaced by stock photos many years ago (see: photoshopped book covers). The job opportunities at the moment are the same as it was before. You just need to be really, really good to stand out above the crowd to get a gig.

[Shillitron]

Just to Clarify Valves position on AI.

https://www.eurogamer.net/valve-says-ai ... t-on-steam

As the legal ownership of such AI-generated art is unclear, we cannot ship your game while it contains these AI-generated assets, unless you can affirmatively confirm that you own the rights to all of the IP used in the data set that trained the AI to create the assets in your game,

Then, a week later:

we reviewed [Game Name Here] and took our time to better understand the AI tech used to create it. Again, while we strive to ship most titles submitted to us, we cannot ship games for which the developer does not have all of the necessary rights. At this time, we are declining to distribute your game since it’s unclear if the underlying AI tech used to create the assets has sufficient rights to the training data.

Follow Up:

We know it is a constantly evolving tech, and our goal is not to discourage the use of it on Steam; instead, we’re working through how to integrate it into our already-existing review policies. Stated plainly, our review process is a reflection of current copyright law and policies, not an added layer of our opinion. As these laws and policies evolve over time, so will our process

In the meantime it says that it will refund the normally non-refundable app submission fee in cases where this in-progress policy is the deciding factor.

It sounds the same but really Valve is really just trying to insulate themselves from litigation.
These things are currently in pending cases..

--

Which we may have an update on soon.
If you recall, some ambulance chaser and 3 rando SJW trash artists rushed to sue Stability & Midjourney "for the honor of all artists" but really just wanted a quick payday.

Currently a motion for dismissal is sitting on the table.
(With StabilityAI doing a bit of a humble brag)



From the Document

Which states what we already knew, three random shit artists tried to get a quick payday but none of them could cite a single instance of how their copyrighted works were even infringed at all.

ARGUMENT
A. Plaintiffs Concede that McKernan and Ortiz’s Copyright Claims Fail.
McKernan and Ortiz do not appear even once in the body of Plaintiffs’ Opposition. Instead,
they are relegated to a footnote: “Plaintiffs McKernan and Ortiz concede that they did not include
any material that was registered at the time of the filing of the Complaint.” (Opp. at 4 n.3.)
Plaintiffs present no reason why McKernan and Ortiz’s copyright claims should not be dismissed;
and Plaintiffs do not suggest that such dismissal should be without prejudice. Accordingly, Counts
I and II should be dismissed with prejudice with respect to McKernan and Ortiz.

B. Andersen’s Copyright Claims Also Fail.
Turning, then, to Andersen, the Stability Defendants have provided several, independent
reasons why her copyright claims also fail. (Mot. at 6–8.) In response, however, Plaintiffs attempt
to leverage vague, group pleading to support a single plaintiff’s—Andersen’s—claims. But
pleadings about “billions” of unidentified others’ works cannot save Andersen’s copyright claims;
her claims, too, should be dismissed with prejudice

--

We have a new contender as well.
It really feels like law firms are rushing to hit AI companies and be "first in line" to get a big pay day if they can.. Even if they are making a bunch of retarded mistakes in the process.

[wndrbr]

(timestamp 25:23)

Adobe asks the senate to outlaw copying an artist's style!

(Adobe also developed a surveillance tool that can be used to determine if "copyright infringing" AI models were used)

Adobe's CAI cryptographic metadata system is fundamentally a surveillance tool. It's designed to permanently record a file's origin and edit history, and (optionally) the identity of the file creator. It's an important component of the new AI content moderation systems.

CAI's metadata tech has been in development for years, and the original stated intent was to use it to help stamp out "disinformation" online. Right now, ai concerns (copyright and deceptive imagery) are being used as justification for implementing this technology.

C2PA published two papers outlining how their surveillance metadata tech could ruin our lives. It includes an "optional" feature to sign content with a government-issued digital ID. [1] [2].

C2PA metadata tech is designed to work in concert with Microsoft AI content moderation tools to scrub the net of "disinformation". It is being implemented at the hardware level, including by camera manufacturers.

It's likely that C2PA metadata can be easily stripped from a file, however:
- The intent seems to be to quarantine content without C2PA as "untrusted" (i.e. potential disinformation)
- Metadata can sometimes be recovered (centralized database?)

They want to control what information we "eat." They call C2PA a "nutrition label" for information. A centralized authority will define what information is "good" and which is "bad" to eat.

US Department of Defense (DARPA) is very interested in C2PA's potential to identify "disinformation." "Erosion of trust in information."

Microsoft is using C2PA metadata tech to create "reputation management systems" for media uploaded online. They envision users leaving feedback on files that are tracked using C2PA so that they can be reported for "misinformation" etc.

I think they want to get artists on board with C2PA tech, since it's initially being introduced in creative software ecosystems.

[rusty_shackleford]

Just to Clarify Valves position on AI.

[...]

tl;dr:
don't be stupid and say your product contains AI content, they can't prove otherwise

[Shillitron]

How do you outlaw an artistic style?
Is "Anime" an artistic style?

This poo is fucking retarded.

[♦ Arcana ♦]

Western nations are going to cripple themselves (again) with excess rulecuckery because they watched too many schlocky sci-fi films from the 80s & 90s while Asian countries keep doing whatever they want to. The eastern century is upon us.

[General Reign]

"AI art will put artists out of work!" you have it backwards. We already have a society of bad art and no good artists. The reason why "AI art" is a phenomena now is because most modern artists are awful. Academia spent 100 years destroying art. You need to teach people how to create good art again before you can complain about the machine creating better art. Good artists are very rare and it is very difficult to find them, which is why they are booked months in advance and are expensive to contract. That's why the book industry shifted to using photoshopped book covers. The book industry has accepted that it is okay to not have art. AI art (remixing preexisting graphics) is functionally no different than photoshopped book covers, which we already have now.

The whining about AI art now is superfluous. Most of the paying gigs have either been using it already for composition purposes or were replaced by stock photos many years ago (see: photoshopped book covers). The job opportunities at the moment are the same as it was before. You just need to be really, really good to stand out above the crowd to get a gig.

In the years to come as digital replaces more traditional methods, people like the late Bob Ross will be witnessed jumping from the top of tall skyscrapers in mass. It's all digital now. I can hardly find actual paintings online that were not painted a long time ago. It sucks. On one hand digital is easier to use due to layers, on the other I do not like the way it looks, and the intangible nature of it makes me hate it.

[Segata]

This is horrible and problematic and...

[General Reign]

This is horrible and problematic and...

I wonder how many people responded with "nice tits slut" haha.

[Tweed]

People will do what they've already been doing and grift money in the form of donations for either making and improving models or using AI generated content to make stuff people actually want.

[General Reign]

[Red7]

Genie is out of the bottle, they can't stop it, they never could. The people who adopt AI assistance into their work will be the ones who survive.
For now.

i dont think thats the point (i mean ai related job, not that they cant stop technology). short term sure, but deflation/technology and idea of job markets adn inflationary debt expansion are contradicting.

u simply dont need slaves anymore who justify their existence by selling their time hence depopulation push after overpopulation push earlier for cheaper soldiers and labour.

the proper solution is owning ai and owning part on planetary resources/hence benefiting from increased productivity rather than relaying on lack of automation for employment. in that matter niggers got it right; we are not meant to work and technology should give us free shit. but for that u cant be retarded nigger who will sell his shit to jews or be force liquidated for debt or whatever.

People will do what they've already been doing and grift money in the form of donations for either making and improving models or using AI generated content to make stuff people actually want.

no they dont. thats the point, advanced ai will not require team of geeks to make it run well and u wont need to hire anyone to utilise it as you wish. bots run by indian scammer or state agency will be more obtrusive and add to the fog of war but all the art bullshit grift will get demonetized just like voice acting should be already demonetized. some idiots will pay for shit u can get for free, sure, but those idiots will get annihilated soon.

[Shillitron]

Legal Update

(Grifters vs Stability AI)



TL;DR Summary
- Three Artists (Grifters) Attempted to Sue DeviantArt, Midjourney and StabilityAI
- Kelly McKernan and Karla Ortiz cases have been dismissed because and I quote: "McKernan and Ortiz, did not actually file copyrights on their art with the U.S. Copyright Office." - Paint Pigs Not sending Their best..
- Sarah Andersen's case against Midjourney & DeviantArt dismissed (See Below Quotes)
- Sarah Anderson likewise still has to prove her copyrights were infringed by stability in an amended claim, she also has to provide evidence of "how Stable Diffusion — a program that is open source, at least in part — operates with respect to the Training Images" - TL;DR Sarah could still have a case but everything filed up til now is considered trash and she's starting from scratch.


One interesting tidbit (Full Quotes Below)

- The Judge seems to have made statement essentially summarized as:

In other words — because AI image generators reference art by many different artists when generating new imagery, unless it is possible to prove that the resulting image referenced solely or primarily copyrighted art, and is substantially similar to that original copyrighted work, it is likely not infringing of the original work.

AKA - This case could actually set precedent that protects AI generate images as fulfilling the transformative requirement to be derivative and protected. So even if Stability eats shit for their LAION training set.. all the users of Stable Diffusion are perfectly in their right to keep on trucking.


Reference Articles

Reuters Biased Article on The Dismissal
(They donate half the article to sucking off the lawyer who basically had his case thrown out for being shit, instead of covering what the Judge said)

U.S. District Judge William Orrick dismissed some claims from the proposed class action brought by Sarah Andersen, Kelly McKernan and Karla Ortiz, including all of the allegations against Midjourney and DeviantArt. The judge said the artists could file an amended complaint against the two companies, whose systems utilize Stability's Stable Diffusion text-to-image technology.

Orrick also dismissed McKernan and Ortiz's copyright infringement claims entirely. The judge allowed Andersen to continue pursuing her key claim that Stability's alleged use of her work to train Stable Diffusion infringed her copyrights.

The same allegation is at the heart of other lawsuits brought by artists, authors and other copyright owners against generative AI companies.

https://venturebeat.com/ai/midjourney-s ... continues/

More Case Dismissal Context,
Mainly around the Judges thoughts / feedback before dismissal that Reuters didn't bother including:

Orrick spends the rest of his ruling explaining why he found the artists’ complaint defective, which includes various issues, but the big one being that two of the artists — McKernan and Ortiz, did not actually file copyrights on their art with the U.S. Copyright Office.

Also, Anderson copyrighted only 16 of the hundreds of works cited in the artists’ complaint. The artists had asserted that some of their images were included in the Large-scale Artificial Intelligence Open Network (LAION) open source database of billions of images created by computer scientist/machine learning (ML) researcher Christoph Schuhmann and collaborators, which all three AI art generator programs used to train.

The Judge continues:

“The other problem for plaintiffs is that it is simply not plausible that every Training Image used to train Stable Diffusion was copyrighted (as opposed to copyrightable), or that all DeviantArt users’ Output Images rely upon (theoretically) copyrighted Training Images, and therefore all Output images are derivative images.

Even if that clarity is provided and even if plaintiffs narrow their allegations to limit them to Output Images that draw upon Training Images based upon copyrighted images, I am not convinced that copyright claims based a derivative theory can survive absent ‘substantial similarity’ type allegations. The cases plaintiffs rely on appear to recognize that the alleged infringer’s derivative work must still bear some similarity to the original work or contain the protected elements of the original work.”

https://www.hollywoodreporter.com/busin ... 235632929/

Just more Judge commentary. Dismissal of cases filed against DeviantArt and Midjourney (who uses Stable Diffusion to offer a paid subscription service)
This is going to make it very difficult to go after people who generate / use ai art.

In his dismissal of infringement claims, Orrick wrote that plaintiffs’ theory is “unclear” as to whether there are copies of training images stored in Stable Diffusion that are utilized by DeviantArt and Midjourney. He pointed to the defense’s arguments that it’s impossible for billions of images “to be compressed into an active program,” like Stable Diffusion.

“Plaintiffs will be required to amend to clarify their theory with respect to compressed copies of Training Images and to state facts in support of how Stable Diffusion — a program that is open source, at least in part — operates with respect to the Training Images,” stated the ruling.

[Shillitron]

Judge reaffirms my lolyer position from months ago whilst bickering with Glowie Orange.

What is the definition of "Transformative Use".

The LAION Training set is over 100 TB's in size. The checkpoint (used to guide AI) is 2 Gigs in Size.
Every unique image that is trained on - essentially condenses down to a single byte's worth of influence of the diffusion model. The actual image exists nowhere in the resulting checkpoint.

But to say that AI is "Collaging images together" is wrong on a technical level.

It's also hard to say AI is not transformative

[rusty_shackleford]

@orinEsque do you use your own GPUs or do you rent GPUs to use?

[orinEsque]

@orinEsque do you use your own GPUs or do you rent GPUs to use?

Using my pretty 3080. I'd rather avoid an online trail.

[rusty_shackleford]

You know, about 99% of the future is gay bullshit. But at least one thing wasn't a lie. We did get, in some form, rather advanced AI. And I can't wait for my own personal, local AI assistant.

[rusty_shackleford]

SD img2img works pretty well


Would definitely benefit from a GPT-style interaction tho.

[Shillitron]

You know, about 99% of the future is gay bullshit. But at least one thing wasn't a lie. We did get, in some form, rather advanced AI. And I can't wait for my own personal, local AI assistant.

Run local models.. they are as good as ChatGPT.. the only limitation is they can't hold as large of a context (memory) so they get a little forgetful.
Although if you have a good video card you can get up to 2000ish tokens. Maybe more?

best part is they are uncensored and your conversations aren't saved in some san fran database.

[Shillitron]

SD img2img works pretty well

2021 called.. SD img2img is shit. ControlNet.

[rusty_shackleford]

they are as good as ChatGPT.

No, they're not. Unless you're referring to GPT 3.5.

[Shillitron]

they are as good as ChatGPT.

No, they're not. Unless you're referring to GPT 3.5.

Yea your right.. the newest GPT is amaz-

[rusty_shackleford]

Bard output is as good/better than GPT4 now with its latest model update, and afaict it's free to use. It's also significantly faster to generate output, the web browsing integration isn't shit/much faster, and generates 3 outputs at a time to pick from.

OpenAI is going to need to do something to not hemorrhage subscribers.

[Ranselknulf]

Now we don't need immigrants to clean our toilets.

[Rand]

I trust the robot more than any human.
It will do the job it's been programmed to every time (barring equipment failure).

[WhiteShark]

Bard...

[Wretch]

Holy war against the machines.

[Emphyrio]

[Segata]

Why do I get the impression that these "players" urging others to not buy this game are journalists and a few reddit trannies?

[rusty_shackleford]

Bard output is as good/better than GPT4 now with its latest model update, and afaict it's free to use. It's also significantly faster to generate output, the web browsing integration isn't shit/much faster, and generates 3 outputs at a time to pick from.

OpenAI is going to need to do something to not hemorrhage subscribers.

Main critique is that it seems to not be very random at all. Can't tell if it's a lack of training data, or the knobs are simply not tuned right. Generates a lot of the same results if you try to do anything creative.