AI-Created Art Isn’t Copyrightable, Judge Says in Ruling That Could Give Hollywood Studios Pause

A federal judge on Friday upheld a finding from the U.S. Copyright Office that a piece of art created by AI is not open to protection.

"In March, the copyright office affirmed that most works generated by AI aren’t copyrightable but clarified that AI-assisted materials qualify for protection in certain instances. An application for a work created with the help of AI can support a copyright claim if a human “selected or arranged” it in a “sufficiently creative way that the resulting work constitutes an original work of authorship,” it said."

Thaler was appealing this, and his appeal was denied.

mindbleach,

Copyright exists for a reason.

Creation has always been harder than copying - retelling a joke you heard takes little talent. Printing books is easier than writing books.

Copyright is a monetary incentive for new difficult works.

If you can push a button and a new work comes out… the question of who made it is completely fucking irrelevant. Only human effort deserves this protection.

snooggums,
snooggums avatar

Retelling a joke is an actual skill, because some people can do it well and others will ruin it by how the tell the joke.

Copying the exact text if a joke is not.

Copyright is only an incentive for new works if it is limited in duration.

paddirn,

Should photography be copyrightable? If you can just push a button and a new work comes out, does that really deserve protection? I understand it can take more effort than that when you count all the prep work, but at the end of the day, you’re still just pushing a button to create the work.

Quasi,

That’s like saying “All you do is push ‘Render’ and ‘Toy Story’ comes out.”

ThunderingJerboa,
ThunderingJerboa avatar

That is a great oversimplification. I mean it is a fair point Photography is taking images of the natural world. In your example, there is far more texture work,modeling, etc for it. While Photography those kind of things were already done for us, we are merely picking the angle, maybe influencing the lighting, etc. This isn't to belittle Photography since they even acknowledge there is more work to it then just click button. I think we are just seeing people belittle a medium for the barest min requirements of the medium. I wouldn't call taking a selfie of yourself doing a duckface grand art to it even though its photography but I feel people are placing "AI art" in that place where the worst of the worst is stated to be the whole medium, when merely "AI art" is just another tool in the arsenal of real artists.

Quasi,

Oversimplifying it was my point.

In every single art discipline there are ways of creating with very little effort. You can point an iPhone at your nephew and create a photo, but you can also spend weeks stalking an animal for the perfect photo, or a natural phenomenon that rarely occurs. You can arrange the composition of a portrait, spending hours getting the lighting right. That’s ignoring the post-processing that can consume a crazy amount of time.

People think that because the technical side of photography is more accessible than in the past, there’s very little to the art.

But consider a random stranger’s phone. What are the odds that there’s a single photo on there that you would enjoy looking at, let alone having a print on your wall? Probably very low.

But then compare that to a professional’s portfolio. There’s going to be a huge difference, because a professional has the eye for a good photo. And people will happily dismiss their work because it only took 1/64th of a second to capture, and they’ll ignore the insane amounts of work that went into honing their ability.

As I said, every single discipline has low-effort works. The very idea that photography is by it’s nature so low-effort that it’s not worth copyrighting is incredibly short sighted.

ThunderingJerboa,
ThunderingJerboa avatar

As I said, every single discipline has low-effort works. The very idea that photography is by it’s nature so low-effort that it’s not worth copyrighting is incredibly short sighted.

I think their point was this case seems to be biased against new technology. We don't put into question that photography is copyrightable since its been an established art form. Even someone who just takes random pictures on their phones are able to copyright their work because even at the end of the day photography is a bygone technology and many of the rights they had for them were developed decades if not a century after the advent of the technology. Many of the copyright protections that photography has these days date back to 1988's Berne Convention, obviously there were earlier laws on the subject but they did vary quite a bit.

The very idea that AI Art is by it’s nature so low-effort that it’s not worth copyrighting is incredibly short sighted.

Ragnell,
Ragnell avatar

The precedence in it ACTUALLY involves photography, where a monkey took a picture after someone had left a camera by a group of monkeys. A picture taken by a human would be copyrightable, the picture taken by the monkey, not so.

Consider it this way. IF you are putting the same amount of effort you would put in that you would put in to commission a human artist to make work for you, then it's not yours.

If you are putting in as much effort as you would make to fully collaborate with another human, then it's yours.

What this basically does is say that corporations and people can't use AI to commission art or work-for-hire and have it copyrighted. It explicitly says that if a human uses AI as a tool they can still copyright.

ThunderingJerboa,
ThunderingJerboa avatar

It explicitly says that if a human uses AI as a tool they can still copyright.

The mere fact a human made a "prompt" in theory should be a good enough category for it be copyrightable but apparently not. As I said, having someone make a duckface selfie is automatically a copyrightable work, which requires so little effort or human element to it that it is insulting. Again I am fine with this result if the parameters of "Humans using AI as a tool" is drawn well enough in the future, right now that line is very ill defined. The problem is people are scared of this technology and there are reason to be if its implemented poorly it can hurt a great many people but we are clearly seeing people draw very hard lines that are unrealistic and these companies will do everything they can to save money. So these kind of ruling mean very little in the grand scheme since no shit Disney or whatever company probably was never going to just sell a raw collection of renders as a product besides the most budget of budget studios.

Ragnell,
Ragnell avatar

I think you're not giving people who know how to angle a selfie credit, but that's a tangent.

The law needs to start somewhere, and "you can't use AI for work-for-hire and have the copyright" is actually a pretty good place. Because Thaler wasn't claiming the copyright for himself, he was claiming the AI as the creator and himself as the owner for "Work-for-hire", kind of like how Siegal and Shuster created Superman but DC got all the money and the ownership for it.

That is precisely what the corporations trying to screw over the WGA wanted to do, have scripts from AI and themselves as owners for "work-for-hire."

Now that particular use case is struck out, and there's still room for people who use AI as tools to claim copyright personally. I think the line may have to do with how detailed the prompt is. Like, I can say "Draw a man punching another man", or I can describe in detail the man, the posing, the other man, their facial expressions, the angle...etc... "Draw a man punching another man" is way lower effort, I think, than even taking a picture of myself, because I'm not even choosing an angle. But if there's a detailed scene set, then I think I would have enough to argue with the copyright office. I'd save my prompts if I were going to test this.

lol3droflxp,
lol3droflxp avatar

The work is in finding the right angle, subject, lighting, often post-processing, cropping, etc

wagesj45,
wagesj45 avatar

Just like the work is in crafting the right prompt, finding the right models, finding the right custom training data, and then putting it all together.

lol3droflxp,
lol3droflxp avatar

Maybe

hglman,

Just like it takes a lot of effort to make an AI that does what you want.

Ragnell,
Ragnell avatar

That's a WHOLE other can of worms. The guy who made the AI is not necessarily the guy who uses it, and the way it's been made may have been using the efforts of other people without compensation.

hglman,

Hard to say that AI using others’ work and a photo of others’ works is different. It would make take taking pictures of anything remotely urban illegal.

No, the person who made the AI is analogous to the person who manufacturers the camera. The person getting the AI to make some product is the same as the photographer.

What makes AI different here?

Ragnell,
Ragnell avatar

@hglman

Level of control. Everything in photography is a deliberate choice. Subject, framing, angle. Even a selfie is trying to get just the right angle. You see it with your own eyes, and select the moment you want to capture.

Sure, a straight-on photo of the Mona Lisa is not gonna be copyrightable, it's like a Xerox of the painting. But take this image of the Scranton City Hall for example.

https://assets.simpleviewinc.com/simpleview/image/upload/c_limit,q_75,w_1200/v1/crm/lackawannapa/Downtown_scranton_City_Hall_Night_02_Mike_Lehman_4993D4CC-5056-A36A-07E0DD612704117C-4993d3dd5056a36_4993d528-5056-a36a-07c1ebde31644933.jpg

The photographer had to choose the time of day, the exact right spot to stand, the angle to hold the camera, and the overall tone of the picture. It is distinctive from other pictures of Scranton City Hall taken at other times of the day, or from other vantage points.

It is a distinct, deliberate image meant to convey a certain image.

With AI, you enter a prompt and the computer does the framing, angling, pacing, focus...etc.. You don't have any input on the actual process if you are just entering a prompt. Now, you do if you are creating digital art using AI as a tool in your process to do smoothing and such (which is considered human authorship by the copyright office), but for things like ChatGPT and Midjourney where you just enter a prompt? It's not your arrangement unless you choose from a selection of AI outputs and put them together as a series or collage (also considered human authorship).

Scranton City Hall itself is a work designed by Edwin L Walter and Frederick Lord Brown.

If you go in person and take a picture of Scranton City Hall, the photo is your work of art. It is your copyright.

You can create an image of the Scranton City Hall using AI to help with the details, making the choice whether to keep or roll back the AI changes as you work towards the final product, and it will be your art. it would be copyrighted to you.

But if you type "an image of Scranton City Hall at night" into an AI prompt, you are getting a work where the program has determined what the angle, mood, tone, and vantage point is. That's not your art. It is the AI's an AI, like the monkey who took the photograph, is not protected by copyright law and therefore the image is not copyrightable.

What I'm saying is a teen girl angling her wrist to find the exact right angle where a blemish on her forehead and a little extra fat on her cheek is invisible is making more deliberate choices about the final image than a guy who puts "image of a teenaged girl making a duckface" into an AI prompt, and she deserves to have her image copyrighted.

hglman,

Your comparing simplistic use of AI to professional photograhy. Thats not a compeling argument.

Level of control is fairly vauge point. If you use auto focus do you no longer get to copyright a work? Do you need to name your goals of the photo ahead of time so we know you didn’t capture something by accident ? What about the fact that the camera holds all the parts together for you, shouldnt you need to personally align the lesne and film as well as move the shutter. Oh what if the shutter speed is faster than whats humanly possible?

How do we test that a tool has enough input complexity to make its output human generated enough?

Ragnell,
Ragnell avatar

The ruling is ABOUT the simplistic use of AI. Right down at the bottom it says that a human using AI tools is still a human author. Thaler tried to copyright the AI as the author, and himself as owning it due to work-for-hire. The Judge said no. Because if you just put a prompt in and get art out, you are not a creator.

Your examples are all of a human using automated tools to make art when the ruling specifically states that a human using automated tools is an author. The work is only non-copyrightable when AI is the author.

The precedent is of a freaking monkey taking a picture, for asimov's sake. Honestly, photographers do not deserve the shit they're getting piled on them in this thread.

Ragnell,
Ragnell avatar

@hglman Also, I'm hurt. I gave you a distinct example of the use of AI as a tool to assist an artist in that comment so that I could COVER the use of AI as a tool rather than having an AI as the author and you say my argument only covers "simplistic use of AI."

Redonkulation,

I really don’t understand why it feels like so many people are desperately ready to let AI tools annihilate the arts, and defend it so viciously.

Sure it’s fun now, but in reality all it’s going to do is completely enshitify graphic design, photography, music production, marketing and commissioned arts as career options. All of these systems are so valuable because the largest corps in the world are drooling at the idea of cutting all the human cost of any creative fields they engage with.

Why are people who aren’t in those positions of power so excited for AI to completely devalue the arts? It’s the most human thing we have and people are so ready to just piss it away to computers.

Fuck generative AI.

mindbleach,

Remember when Terragen got really good, and nobody painted landscapes anymore? No? Well there you go.

This is going to go terribly for media giants, because if they don’t need people to make stuff, neither do you. You, right now, can push a button and get a deluge of whatever images you can describe. They’re not great. But the technology has existed for something like eight months, and any shortcoming people can spot is feedback that makes it better.

And stop trying to ruin the term “enshittification” when this has nothing to do with monopsony.

In any case:

Demand 30-year copyright, no exceptions. We need to restore the public domain.

Mandatory licensing is unavoidably necessary with streaming’s petit monopolies. All media platforms should serve all media.

Any humans left in the loop who aren’t getting a cut of their runaway successes have been robbed, and the people responsible belong in jail.

Technology is never what’s wrong.

Redonkulation,

Sorry it’s hard to tell from your comment, do you agree or disagree that AI generated content should not be able to by copyright?

mindbleach,

…

Redonkulation,

If you’re gonna have a cunt opinion, at least be able to articulate it when asked.

mindbleach,

I was crystal fucking clear, at length.

You are illiterate.

Ragnell,
Ragnell avatar

The answer I see the most when I ask how AI has changed the world has been "Art is now available to the masses and not just a small group" meaning it's people who think that making is something only select, talented people can do and not something that anyone can do. I suspect a certain amount of jealousy, and that they don't understand that accomplished artists have spent years honing their skills, so they cheer AI as making them able to produce artwork when really it's just another form of asking someone else to do art for you.

Redonkulation,

I can empathize with the feeling of thinking “gosh I could never do that, your work is amazing”, and I think especially now with social media, people only see the final product, skipping the countless hours of work and years of dedication. Kinda like the old “man your camera takes amazing pictures” thing.

As a person that is on that journey with multiple artistic disciplines, some of my lowest days have been when I have struggled to improve my skills. Its depressing and it hurts, but with those downs comes the peaks of making genuine improvement and progress. I can ride that feeling of a single success for months. Its dramatic, but I worry for people younger than me being completely disconnected from the arts and the personal journey that is so important.

Ragnell,
Ragnell avatar

That's why it's so important, I think, to let people SEE the creative process, and let people SEE your earlier work while you were learning.

FaceDeer,
FaceDeer avatar

Getting an AI to make good art takes a lot of human effort, at least with the current generation of AI art generators. But fortunately that appears to be accounted for:

AI-assisted materials qualify for protection in certain instances.

I imagine it's no different from creating a piece of art by starting with a copy of something conventional in the public domain, such as the Mona Lisa. If you apply enough of your own work to it it becomes yours.

Ragnell,
Ragnell avatar

Yeah, it's like... Well, if you put a prompt in Midjourney and it generates a picture, that's commissioning art from an AI. If you are using digital drawing tools and use some AI for finishing touches, that's making art.

Or if you commission works from midjourney, then arrange them into a cohesive whole like a collage, then the collage is art by you but the individual pieces are art from AI. If the studio puts a prompt into an LLM and gets a script, it's not copyrightable. If they pay a writer who uses an AI to correct the grammar on a script or search for errors, that's copyrightable.

It's an easy distinction until you get into AI-generated performances of people's likeness. That'll be... interesting.

elscallr,
elscallr avatar

There's a lot of middle ground between this two extremes, though. Suppose you iterate on your prompt again and again and again to exact the image you had intended. Does the end result constitute art? Would it ever? What about the prompts? There was difficult effort, but no "digital brush" on a "digital canvas" (so to speak).

Ragnell,
Ragnell avatar

Yeah, there is. The copyright office is going to just have to make a determination in each case for a while.

ThunderingJerboa,
ThunderingJerboa avatar

But with things like Stable Diffusion you have tools that are like inpainting and outpainting where you can brush out/in details you want changed. I think this is a far more interesting question since in your own its just a raw output from AI just with a revamped prompt which the court stated can't be copyrighted. Unless what you meant was you first run a Txt2img pass then you use img2img with some minor changes like drawing something or drawing things out or changing colors, would that then quantify as a "copyrightable work" since then you do have a more human element to direction.

wolfshadowheart,
wolfshadowheart avatar

In-painting really makes it difficult. What's the difference between in-painting an image (that is to say, AI generated texture) vs photoshops Clone Stamp?

wave_walnut,
wave_walnut avatar

If a capitalist had an AI that learned Syd Mead's work output many Syd Mead-like illustrations for publication, he would be sued by fans to protect Syd Mead's copyright. But if the same thing were done by an AI that learned the work of a new artist who admired Syd Mead and honed his skills, who would protect the new artist?

sunbeam60,

The argument about a photograph being copyrightable, even if made by a machine, because a photographer has chosen pose, lighting, lens settings etc., becomes extremely elastic as AI outputs become more steerable: “put a lamp here, turn the light on, make it cosier” etc.

Is it actually the prompt-chain that is copyrightable?

It gets very confusing, very very fast.

Ragnell,
Ragnell avatar

Funny, because photography is actually the precedent on this. A monkey took a picture, it was not copyrightable.

I'd advise you to keep a record of your creative process here, because it may come down to how many prompts you used to steer it.

apparentlymart,

It’s nice to see this starting to get tested in court.

I doubt this will really upset “Hollywood Studios” too much, though. They are unlikely to be creating entire productions using ML techniques, and instead using it for smaller parts of an overall production. It seems like e.g. generating a voice or image for one part of a film would not invalidate copyright on other parts of the work or of the overall work taken together.

Film studios also rely on other non-copyright protections such as trademarks to dissuade derivative works.

I think the bigger test will be: is it copyright infringement to use a work as part of a training set for a model? It’s all very well saying that the output is not copyrightable, but there’s still a big question about the input.

gullible,

What is there to prove that a piece of artwork was or wasn’t made by AI? Within a few years, it will be nigh impossible to tell when AI helped or wholly created a piece, and this will effectively stop no one. It’s a bandaid on a hemorrhage. It’s a very helpful bandaid, but by fuck can someone grab the sutures already?

FaceDeer,
FaceDeer avatar

What would the "sutures" be, in this analogy?

gullible,

That’s a difficult question to answer and one that will inevitably piss off many. Mandatory detectability is the easiest answer, but the vaguest- ideally with information about used artworks. Severe fines for training on copyrighted works is another route. Vast and continued investment in creating a public domain art library to pull from would ease the issue for artists. I don’t use AI for art and I don’t legislate so I’m not the one to ask, all I can offer is bandaids as well.

Ragnell,
Ragnell avatar

Well, computer forensics IS a thing. Computers keep a record of everything done on them, and if it comes down to a lot of money at stake and a lawsuit then those computers can be looked at.

gullible,

What, a full department equivalent to the IRS to constantly audit art and its process? That could work. It’d be pricy, but it could work.

Ragnell,
Ragnell avatar

It would create jobs.

DrYes,
DrYes avatar

I didn't read the article but does it say anything about detectability? If I claim I created it who can prove an AI did?

moosetwin,
@moosetwin@lemmy.dbzer0.com avatar

I didn’t read the article but

DrYes,
DrYes avatar

So I read the article and the lawsuit has nothing to do with my question. It's more about precedent and actually finding out what the laws are if everything is as claimed.

Still think detectability is an interesting question.

Ragnell,
Ragnell avatar

Yeah, but detectability isn't a new question, is it? It's just a twist on the old question of "Did someone else create it other than the guy who claimed it?"

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • tech
  • DreamBathrooms
  • magazineikmin
  • everett
  • InstantRegret
  • rosin
  • Youngstown
  • slotface
  • love
  • khanakhh
  • kavyap
  • tacticalgear
  • GTA5RPClips
  • thenastyranch
  • modclub
  • anitta
  • mdbf
  • tester
  • Durango
  • ethstaker
  • osvaldo12
  • cubers
  • ngwrru68w68
  • provamag3
  • normalnudes
  • Leos
  • cisconetworking
  • megavids
  • JUstTest
  • All magazines
    •