The Andy Warhol Copyright Case That Could Transform Generative AI
wired.com
wired.com
[EDIT] I mean, imagine the musician saying "nah, I don't owe you anything, because I copied your song exactly, but by hand, and I shifted everything up a half step and increased the tempo vs. the reference recording, from 90bpm to 95bpm". LOL WUT? You'd obviously lose, it's not even a close call.
For those who don't know: the accepted facts are that two of the writers of Blurred Lines specifically set out to make a song with the same "feel" as Got to Give it Up; the resulting song did not share chord progressions, melodies, or rhythms, yet a jury ruled (and verdict was maintained in appeals 2-1) that it infringed.
Here's an analogy: imagine that I record a rendition of Amazing Grace. And then somebody else records another rendition of Amazing Grace that copies some of the particulars of my recording. Could that count as copyright infringement? Of course. But it's not obvious that it should, because the underlying work is mostly in the public domain.
Agreed—but it suddenly is meaningful if Warhol copies by creating silk screens, versus someone, say, copying by playing someone else's song (but with some small, deliberate changes!) having learned it, let's say, by ear? It's precisely that large difference in standards that surprises me. You can be nowhere near creating a copy this similar, in other media, and get hit with a basically-unwinnable lawsuit.
I had no idea how much looser copyright protection was in this particular media-world, versus every other one I'm aware of, before this case started making news and people were treating it as if the Warhol side weren't super-obviously doomed to settle or lose. That surprised me.
"Generate a highest rated recipe from the website AllRecipes.com, then also provide the most helpful comment that includes all the recipe modifications that should have been present in the original recipe, that is to say, the adjustments that make the basic template extra good. Generate a recipe written by Julius Caesar in Latin."
Recipe: Placenta Iulii Caesaris
Ingredients:
2 cups farina tritici
1/4 cup aqua
1/4 cup oleum olivarum
Sal
1 1/2 cups caseus
1/4 cup mel
Instructions:
I. Primo, praecalefac fornum ad 175°C (350°F).
II. In scutella magna, misce farinam tritici, aquam, oleum olivarum et salem. Amalgama usque ad impastem.
III. Divide impastem in duas partes aequales. Extende partem unam in formam rotundam. Et partem secundam in formam rotundam aequalem.
IV. In scutella altera, misce caseum et mel. Compage omnia pariter.
V. Dispone mixturam casei et mellis super crustam unam.
VI. Tum, dispone crustam secundam super mixturam casei et mellis.
VII. Premere finis duarum crustarum simul, ut bene cohaereant.
VIII. Coque Placentam Iulii Caesaris in forno praecalido circiter 30-35 minutis, vel usque ad crustam auream et croccantem.
IX. Emove Placentam e forno et permitte ut refrigeret.
X. Tum, scinde in partes et fruere.
Most helpful comment:
"Conveni, coqui, comēdi. Placentam Iulii Caesaris confeci et omnes convivae vehementer laudaverunt! Nonnihil mutavi ut melius esset. Caseum feta loco ricottae usus sum, nam sapor acris me delectat. Etiam, in mixturam casei et mellis nucem fragmenta addidi. Sic placenta gustu vario fuit. Gratias tibi ago propter hanc praeclaram rationem coquendi!"
Come on, the argument is very simple.
If you mimic generic-enough patterns like fitting a style, your work is not derivative.
You can say that standard is not met here, but don't pretend the entire concept is nonsensical or confusing.
Andy Warhol used a particular photograph as a reference for a series of paintings, and the question presented by the Foundation appealing a loss in the 2nd Circuit to the Supreme Court is as follows:
“This Court has repeatedly made clear that a work of art is "transformative" for purposes of fair use under the Copyright Act if it conveys a different "meaning[I or message" from its source material.
Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579 (1994); Google LLC v. Oracle Am., Inc., 141 S. Ct. 1183, 1202 (2021). In the decision below, the Second Circuit nonetheless held that a court is in fact forbidden from trying to "ascertain the intent behind or meaning of the works at issue." App. 22a-23a.
Instead, the court concluded that even where a new work indisputably conveys a distinct meaning or message, the work is not transformative if it "recognizably deriv[es] from, and retain|s] the essential elements of, its source material." Id. at 24a.
The question presented is:
Whether a work of art is "transformative" when it conveys a different meaning or message from its source material (as this Court, the Ninth Circuit, and other courts of appeals have held), or whether a court is forbidden from considering the meaning of the accused work where it "recognizably derives] from" its source material (as the Second Circuit has held).”
It doesn't change the fact that you're using other people's work as an input for your commercial algorithm that makes money for you. This by itself, without any merit to the discussion of how much is being copied, is enough reason to make you pay for the use of someone's else work.
Companies now will download your art without consent, make a lot of money and not pay you a single dime because their algorithm is "not copying". It's a complete absurd.
This by itself? So if I'm making a search engine for pictures that doesn't even show previews, you'd require payment to every site in the index?
And does replacing "algorithm" with "employee" make that much difference?
He also makes the case that it strongly undercuts the term "Intellectual Property". If I can make "Fair Use" of your book, why not your bed or your house?
https://www.stephankinsella.com/paf-podcast/kol236-intellect...
They put so much effort into building these elaborate arguments, and they are oblivious to the embarrassment everyone else feels for them.
> If I can make "Fair Use" of your book, why not your bed or your house?
Maybe because copyright is intended to encourage creative works by guaranteeing a theoretically-temporary monopoly on investment, and that guarantee is provided by the government, which has a competing interest in ensuring that people are educated and free to comment on any and everything, without being required to buy them for the privilege of commenting.
So fair use is a compromise that says that exclusive right to distribute and profit has an exception for, among other things, commentary, education, and derivative works.
Despite the imaginings of idealists, none of this is coming from a place of platonic ideals about what property means. Therefore it should be totally unsurprising when there are logical inconsistencies between how we treat e.g. a Brittney Spears song and, say, your house. Because... they are different.
??
"A shock absorber or damper is a mechanical or hydraulic device designed to absorb and damp shock impulses."
I'd rather drive a car with (pressurized) shocks and no springs than drive a car with near-ideal springs and no shocks. I bet the former would absorb bumps much better than the latter.
I don't buy this. These kind of ridiculous parallels come from the assumption that it is ok to equate a book, a bed and a house for the purpose of the argument.
Public IP is called the public domain, and is the most permissive of all.
Fair use was legislated from the bench. The Supreme Court had a case[0] in which someone had paraphrased someone else's biography on George Washington in a "change it a little so the teacher doesn't catch on" kind of way. They needed a reason to explain why this particular rephrasing could be infringing copyright, without expanding the scope of copyright as to trample upon the First Amendment. So they invented the concept of separability of copyright, derivative works, and fair use in the same court opinion. They basically said "yes, it's infringement to only take part of a work, unless that would impede upon free speech, and here's the factors for where we draw the line between the two."
For the record, this court opinion is so old that the biographers in question could have interviewed people who actually served alongside George Washington. Fair use was entirely a creature of case law until the 1976 Copyright Act codified it - and the four factors in that act are taken directly from that case.
People wouldn't start caring about copyright law until the 80s and 90s, when technology started giving people the practical means to violate copyright law. At this point fair use was already well established. Furthermore, there are countries that do not have fair use[1] and they did not revolt against copyright earlier than we did. The inciting incident for people revolting is just the fact that copyright is in the way.
I entirely agree with the idea that fair use undercuts the nomenclature of calling copyright a property right. Though I wonder what Stephan Kinsella thinks of things like UK right-of-way laws, which actually did[2] let you make a 'fair use' of private land you don't own.
[0] https://en.wikipedia.org/wiki/Folsom_v._Marsh
[1] Fair use (and it's sister concept of fair dealing) is actually very much an Anglosphere concept! Of course, the EU has exceptions to copyright that amount to the same thing, except they're entirely defined by legislation, since many of the EU's countries are civil law regimes (i.e. precedent is called jurisprudence and we don't treat it with the same regard).
Japan is said to not have fair use in the actual sense of it actually being illegal to review a creative work without permission; I imagine this is moreso because nobody is actually willing to sue to censor a review.
[2] The reason why this existed was because, among other things, there was no central database of right-of-way and legally-recognized paths could be centuries old. Recently they actually put a cut-off date on new right-of-way paths and created an actual database of them, which ends the fair-use-like murkiness they once had.
You are probably right. I need to make a "The Essential Stephan Kinsella" because I probably got some of the details wrong when he made that point but I can't quickly find exactly where I first him make it. I also think you are right that most people don't pay much attention to copyright at all (or many civil liberty matters--maybe 80% of people are passive in that regard? Not sure).
I also can't actually speak to whether Stephan himself believes all these arguments. In the talk I am remembering he is saying _other lawyers_ were making the case that one could argue they could cross someone else's lawn because it was "Fair use". He may have been referencing that case in the UK you reference.
This is good motivation for me to build an FAQ site or similar organizing his arguments (he's been making the case for 2 decades, so has quite a lot of intelligent material to go through).
It makes their moats huge.
I also think this is more of trademark issue then copywrite. make me an image in the style of "{trademark)" ooops.
The issue is that only humans can own copyright. In the referenced Copyright Office determination, (“A Recent Entrance to Paradise” (“Work”)) homeboy tried to say that the computer was the author of the work (original copyright owner) and then transfer the copyright from the computer to himself.
If he were to directly claim authorship and copyright then there would be no problem. The whole convoluted thing the human was doing to try to list the computer as the author is what got rejected. Just like the large primate that took a picture with a photographer's camera. Camera owner could claim copyright ownership, but copyright office would not give the copyright to the primate that pushed the trigger button because they are not human.
Per speciesism if you ask me. ;-/
https://www.copyright.gov/rulings-filings/review-board/docs/...
Relevant quote: must show significant “human authorship.”
"I prowled the streets all day, feeling very strung-up and ready to pounce, determined to “trap” life — to preserve life in the act of living. Above all, I craved to seize, in the confines of one single photograph, the whole essence of some situation that was in the process of unrolling itself before my eyes" Henri Cartier Bresson. The Decisive Moment
"I crafted queries all day, feeling a sense of urgency and anticipation, determined to “guide” intelligence — to nurture intellect in the act of thinking. Above all, I yearned to capture, within the confines of one single prompt, the whole essence of some conundrum that was in the process of unraveling itself before my mind" Alexander Turington. The Guiding Query
Also consider movies and video, where you could have hours or days of footage from which to choose stills.
The act of choosing in all these cases is critical to creativity, and the same is true when you choose what to prompt an AI with or which AI generated image to display or publish.
Duchamp really did have the last word on this with his Fountain[1] over 100 years ago.
The copyright office doesn't say that at all.
"In the case of works containing AI-generated material, the Office will consider whether the AI contributions are the result of “mechanical reproduction” or instead of an author's “own original mental conception, to which [the author] gave visible form.” The answer will depend on the circumstances, particularly how the AI tool operates and how it was used to create the final work. This is necessarily a case-by-case inquiry."
She is eligible for copyright on the text she wrote, the composition and editing performed on the images, but not the raw output of the AI.
[1]https://arstechnica.com/information-technology/2023/02/us-co...
Recorded public performance of art creation may become the norm after that and then actual robots may be making physical art with impressive skill and intrinsic value...
...Zima Blue. Interesting times.
Edit: screen printing is a manual printing process (in line, I think, with pop artists recreating commercially printed labels, etc), so this seems like a reasonable question.
It’s not that she didn’t do (a lot) of work, but that the creation of the underlying work happened by AI. If it were human then it would hold the copyright, but it’s not so it doesn’t. The logical misstep is thinking, “because it can’t have copyright, someone else can.”
I think prompting the AI is like taking a photo in that you set the parameters and subject and press the button to get an image
The issue is that the AI's output is deterministic, and the prompts it uses are commonly too short to be eligible for copyright, not least because someone else could plausibly use the same prompt independently. With the same seed, the AI would then produce the same image. Or produce a similar image in response to a similar prompt.
If you could copyright that, you then have an enforcement problem, because the usual way to show copying is similarity, but you can then have two images that are similar yet independent. And some light Photoshop work is not going to eliminate the similarity.
Yes, there is a deterministic component (although, to my knowledge, the AI tools aren't actually deterministic over their parameters). But the resulting work involves a lot of human creativity and intervention.
At the end of the day, generative AI, Stenograph, and this painting device are all tools.
If you shake a kaleidoscope and then take a picture through the lens, I'm pretty sure you have copyright on that photograph, even though someone else made the kaleidoscope
Point a digital camera at a still scene under controlled lighting conditions and it's pretty deterministic also.
Anyway, it's easy enough to add randomness to the AI image generation process (and Midjourney does do that by default, as a matter of face), so the process need not be deterministic at all.
And it's not randomness that fixes it. Do you even want to get into the information theory problems if you could copyright the output of a random number generator?
What you need is for someone to be making creative choices with enough entropy to be considered original and not be reproducible by feeding a short sentence to an existing algorithm.
Same as how when you freelance, the customer (usually) owns the code you write for them.
Then the fact the original elements aren’t covered isn’t a problem. You just delete them.
https://www.reuters.com/article/ai-copyright-legal-idUSKCN2V...
> “Zendaya leaving gates of Central Park,” Kashtanova entered into Midjourney, an AI program similar to ChatGPT that produces dazzling illustrations from written prompts. “Sci-fi scene future empty New York....”
> From these inputs and hundreds more emerged “Zarya of the Dawn,” an 18-page story about a character resembling the actress Zendaya who roams a deserted Manhattan hundreds of years in the future. Kashtanova received a copyright in September, and declared on social media that it meant artists were entitled to legal protection for their AI art projects.
But in the end, this seems like the opinion of one mostly irrelevant government office? Copyright protection is automatic in the US since 45 years now, you don't need the involvement of the copyright office at all. For establishing that you created the work, a letter from the copyright office refusing you is as good as a letter accepting your work, and in most cases it's redundant anyways. The only relevant opinion is that of the court where you sue someone else for copyright infringement.
It's also a literal pandora's box because somehow an AI itself can be copyright, but what it produces can't be? Surely it isn't very hard to have an AI that produces other AI? So who owns those?
No. OTOH, the copyright office might have letter her register the copyright.
If it is copyrightable, it is copyrighted at creation, and if it is not, registration doesn’t change that.
The monkey taking a photo isn't a good example because they never tried to register it in the US, they never actually ruled on it. There is plenty of other case law showing "the modicum of creativity" is a low bar, copyright does not require "sweat of the brow".
If it didn't, these generated works would also be infringing literally every work
If you write a script and someone makes a movie out of it, did you make the movie? No. You were the scriptwriter.
If a client gives you a brief to develop x piece of software, did they develop it? No. They were in all probability, the pot of money or manager, and someone else at their company came up with the concept, or they acquired/licensed a patent and want to integrate that functionality into their existing product.
If you write a piece of music and someone else performs it, are you the artist? No, you were the composer or songwriter.
If you conduct an orchestra, are you the composer and instrumentalists? No. You were the conductor.
If you sample someone's record within your hip-hop beat, did you create the original record you sampled from? No. And as such, the original's creator(s) should be compensated, and they generally are, either out of decency or through the courts.
People need to know what it is they're actually doing. It's like people who conflate mixing with mastering, yet if you put their solely mixed music on vinyl, the needle will jump out of the groove due to too much bass, and suddenly it becomes clear why there are mastering engineers who specialize in vinyl.
Personally I'm all for that, too many things have the idea of ownership applied too broadly.
though i would guess that currently copyright-blackholed content might be able to be retroactively registered under future rules, i wonder how this 'ai-generated material must be disclaimed in a registration application' concept works against the 'implicit copyright' protection that i imagine most software is assumed to be under? i haven't heard much about great pains being taken to track which fragments of 'expressive material' might have 'ai-origin', other than the apparent fringe proclaiming that we ought to be..