Matthew Allen’s AI art won first prize at the Colorado State Fair. But the US government has ruled it can’t be copyrighted because it’s too much “machine” and not enough “human.”
Why This Award-Winning Piece of AI Art Can’t Be Copyrighted::Matthew Allen’s AI art won first prize at the Colorado State Fair. But the US government has ruled it can’t be copyrighted because it’s too much “machine” and not enough “human.”
Allowing people to copyright A.I generated art could lead to huge issues where someone could just churn out generated images like no tomorrow and throw out copyright claims left and right. It could even lead to situation where you can't really create any art because it's probably something that's already been generated by someone or close to it.
That's what I have been saying. If the courts rule that AI generated art is copyright able. What stops some multi-billionaire from copywriters basically every logical arrangement of words or images or whatever. Heck they would probably even offer to fund employing the copyright office with contractors that they pay for to speed up the process and the government would say it's a good thing because they are saving taxpayer money...
To file an infringement suit they'd need to have paid registration for each work which, even for the exorbitantly rich, wouldn't be remotely feasible for all logical arrangements of words/images. There's probably not even enough space in the Universe or time until its heat death to generate and store all such images.
Even if they did, copyright doesn't protect against against independently created works that happen to be similar or even identical - so they wouldn't be exhausting some limited set of possible works by doing so.
This is more of an issue with copyright law than of A.I. content generation. All you really need to do is create an algorithm that creates images based on every combination of pixels. There were a couple of lawyers who did this with melodies by creating an algorithm to generate every combination of 12-note, 8-beat melodies. One of the lawyers has a TED Talk where he goes into more detail with the issues of copyright laws: https://youtu.be/sJtm0MoOgiU
However with generative A.I you dont need artwork for every combination of pixels. Machine learning seems to be really good at finding patterns in everything we may do or think.
With generative A.I we can use this information to create increasingly more human like output. In terms of art mimic and blend art styles, create new designs based on existing ones etc.
Much more elegant and way cheaper than using brute force algorithms.
This is one reason why copyright/patent law is stupid to begin with. Nothing but a state-enforced monopoly on a slice of all possible information in a category. Imagine if people started copyrighting basic trinomials.
What's stopping billionaires from hiring artists now and just having them slave away at 1632961190 images of whatever they could think of for the same purpose? The art doesn't even need to be good, it just needs to be something you can copyright
Making AI art not copyrightable is probably the best reasonable alternative that we could hope for.
Companies are always looking to cut costs and getting some computer algorithm to churn out endless art without having to pay an artist would be a corporate holy grail. Except that if that artwork then can't be copyrighted and thus monetized (or not as easily monetized), then it ruined or at least lessens their push to replace all their workers with AI.
That's basically what they did for the intro to Secret Invasion on Disney+. They used AI to create it and then touched it up. It still looks like shit, much like the show itself.
Given the presence of stolen artwork in the training data I don’t see why it should be copyright able.
Also award winning? It honestly looks like the kind of liminal mindfuckery most models could output. There’s nothing particularly impressive with the piece.
Eh. I've seen abstract art that people are in awe with throughout my life. And like the uneducated swine I am, I've never thought they were impressive either.
Art appraisers are weird.
Edit: I saw the piece in question. This one is a tricky one, because if a human painted it, it would be impressive. Very nice details. But since it was generated by a machine in minutes..... eh.
I agree completely. I think this is the best solution to the AI replacing human artists problem. Big companies can't use AI to replace humans because if they do, whatever they make will be ineligible for copyright and everyone will be free to rip them off.
This is Allen’s AI-generated artwork, which we can publish without asking him because, as the article notes, it’s not eligible for copyright protections.
Or someone figured out how to reliably tune AI to just happen to generate existing art and tries to use it to bipass/invalidate the copyright on that piece of art.
Drugs are patented, not copyrighted, and handled by the US Patent Office. This is a decision by the US Copyright Office.
Not the same thing, and I would not be surprised if the Patent Office decides drugs designed in part with AI tools can still be patented, while the Copyright Office decides art cannot be copyrighted.
AI are already generating antibody treatments. Companies provide AI with the disease/issue and antibodies that kinda work, then have the AI generate antibodies to fix the disease/issue. The best antibodies are then made in a lab and tested in vitro. However, as somone else noted, antibodies/medications are patented, which is different than copyright. Patents can be done on the process of making the antibody so you patrent the final process of making the antibody, not the AI work to come up with which antibody to make. Source: I attended a Patent Law seminar on this a few months ago.
Tweaks adjustments and corrections need a reference material. The raw AI image needs to be published first for his tweaks to be copyrightable. Anyone wanting to claim copyright on edits should produce the initial uncopyrightable image too.
but then that makes it not his original work and he shouldn't be putting it in competitions.
The first half isn't relevant under copyright law, you can publish edits to old public domain works without publishing the original, this is extremely common even by museums.
I find this an interesting point. My understanding is that AI art just isn't considered enough work on the part of the human creator, presumably because of the idea that you only need to come up with a prompt.
But at the same time, most photographs and videos are copyrightable even if you literally just pointed your phone at whatever without any talent at all. IIRC, the idea of photographs being copyrighted was originally a controversial one, but these days is generally accepted. As long as a human took the photo (and not, say, a monkey, as a famous case found).
Is pointing a phone and clicking a button more of a human contribution than coming up with a prompt? What about if they had to iteratively tune the prompt and mask out parts of the image? In my book, I'd say that's more human contribution than many photographs.
"He didn't create it. He put commands into a keyboard."
"He didn't create it. He pressed the camera trigger."
"He didn't create it. He threw store-bought paint at a canvas."
"He didn't create it. He cleaned some dirt off the wall."
"He didn't create it. He was inspired by gods."
Where you see a categorical difference, I see a qualitative one. AI-generated art can be nothing more than putting words into a blackbox, but it can also be a day-long process of tweaking dozens of parameters to get what you want from the words you put into the box. A child can slather paint onto a canvas without much thought - but that doesn't mean great artists drawing complex, intricate paintings isn't art, does it?
Generative AI is a tool. It can do more than most tools, but still, it is something wielded by an artist.
So...he did something and then by a process something beautiful was created. How is that different from pour painting?
He put words into a box == he just tipped over a can of paint
So he did make it. The tool can't generate anything without something feeding it prompts. I mean technically it can but it will just be random totally incoherent stuff.
Yes, US copyright law requires human involvement to grant authorship. AI generated works are not eligible for copyright and it's unlikely to change unless copyright legislation goes through to yet further restrict copyright.
Did you read the article? In this case he put in quite a bit of work to generate and alter the image:
He sent a written explanation to the Copyright Office detailing how much he’d done to manipulate what Midjourney conjured, as well as how much he fiddled with the raw image, using Adobe Photoshop to fix flaws and Gigapixel AI to increase the size and resolution. He specified that creating the painting had required at least 624 text prompts and input revisions.
And he is essentially claiming that the work should be transformative enough to be copyrightable. Even if the original image is not.
That all makes this case more interesting then a lot of others in the past as it is about AI generation with some human input. Not just someone generating vast amounts of work to find something they like (which likely will never be copyrightable). When this goes to the courts will will help to define the line of how much and what type of alterations are required to claim copyright over the works.
Not all AI work is the same, but I am glad that the copyright office is pushing back on these claims. Putting the burden of proof onto the author that they did have enough input into the work. The big open question ATM is how much input is needed and what that input can look like.
He's lying though. He's pretending the original (wierdly blurry) output was the only AI output, but the details and basically everything else is also AI generated. Nothing is his own skill, brushstroke or even artistic effort/craft, other than prompting the machine-image-generator that he sources the work from.
It goes from the same idea that if you saw someone else's art and made your own art, the copyright of the new art would be yours, and not the one who inspired you.
We've seen edge cases in the music industry (Ray Parker Jr.'s Ghostbusters vs. Huey Lewis and the News' Hip To Be SquareI Want a New Drug ) though in most cases in music, artists routinely borrow each other's elements.
The problem is that AI generative art still requires effort from the programmer (the one who prompted the AI) and went through process of telling it what to do, curating the output, running it back through the AI again to add new elements. If that process is sufficiently long, then it warrants copyright. If that process is insufficiently long, then it challenges the copyrightable merits of artists who make quick art.
In my opinion (removed from the whole AI controversy) is that intellectual property law has been long abused, not by artists and creators but by publishers and studio owners who have used their landlord-esque positions to take control of most art, and then extend their IP rights while denying the public a robust public domain.
And they are the ones that are going to ultimately be hurt by the death of IP laws. Artists will still do art, but for the sake of expression, and then it's a matter of the rest of society making sure they're not exhausted by their day-job (which, according to the strikers in Hollywood, they totally are).
In my opinion (removed from the whole AI controversy) is that intellectual property law has been long abused, not by artists and creators but by publishers and studio owners who have used their landlord-esque positions to take control of most art, and then extend their IP rights while denying the public a robust public domain.
The idea that a strong public domain is beneficial to all culture has been lost. Now we have huge court cases because two different songs use a similar progression of notes. The point of copyright should be to motivate people to create more art, not prevent people from doing it.
Take the whole Under Pressure/Vanilla Ice fiasco. Nobody listening to "Ice Ice Baby" is going to then say to themselves "great, now I never have to listen to Queen again". They are both very different songs in different genres, that use the same guitar riff.
I understand why selling pirate DVD's should be illegal, for recently made movies that are still under print, but transforming a work should not be infringement.
I will shamelessly recycle my comment up the thread:
The modern western conception of art (around which the current legal and economic systems were constructed) is really opposite to the idea that the current AI tools (OR the programmer that used them) should deserve any copyright.
Why ? The concept of art (the modern western one) is that an Art piece is composed of :
1 An Idea
2 A form that is given to that idea by a human artist.
The idea can be given by others, to be constructed by an artist. That is usually a Patron (from where Patreon invented its name) , in spanish Mecenas, that pays the work and directs what idea and even general form it will take (the social practice is called Mecenazgo in spanish, since english has no equivalent word, i will use that ). Example: The Sistine Chapel, which was conceptualized (and paid) by the Catholic Church, including themes and general style, and was given to italian artists like Michelangelo to give the final form, which they drew themselves, with the approval of the church authorities at the end.
The current Ai tools work exacly like the Mecenazgo:
the human person (programmer or not) gives an input (textual, or other), the AI goes brrrrrr, and gives back an image. the person can take ir, or re-iterate the cycle with further inputs until satisfaction.
This is really analogous with how art production ocurred in the Mecenazgo: The human input is the step 1 (an idea), the AI does the step 2 (give form to the idea). The further inputs by humans is analogous to the rough drafts the artist had to give the Mecenas first, the Mecenas described in more details and specifications what themes and forms he wanted, and that repeated until the Mecenas was satisfied with the final form the artist gave back.
The current copyright legal and economic system gives the intellectual property to the ARTIST, that made the step 2, and NOT to the Mecenas of the step 1. Because the Mecenas only had ideas, and the one who made what is considered artistic work, that deserves the legal privilege of IP, is the artist. If all someone did was tell the AI what to draw (i.e. gave an idea, general theme and general form), then the person is only acting as the Mecenas. The MACHINE is doing the artistic work, and since the machine is not a human that deserves the legal privilege, ir should be considered non copyrighted or public domain, just like the picture some monkey took of itself some years ago.
This was not always nor everywhere the social interpretation of WHO is the agent that actually made the art. Before the Renaissance, the western societies considered the Mecenas of step 1 the TRUE ARTIST, because he-she had the idea, and the person that gave form to the idea was considered a low level construction worker like stonemasons, that did not even have its name recorded. If you are wiilling to go back there, we would have to fundamentally change our interpretation of art , artists and rewrite the Sistine Chapel as created by the Catholic Church , and michelangelo is irrelevant.
Copyright is particularly artificial and openly amenable to change to suit the needs of the economy and creators it applies to. So treating this as open ended is probably necessary.
Copyright has for a long time happily provided varying degrees of protection by recognising that one may hold copyright over a work but only over a “thin” or relatively minor aspect of the work.
While there seems to be broader factors involved here regarding the power and market dynamics afforded artists and corporations should AI copyright be protected, there also seems to be plenty of scope to recognise that actual original work can be behind an AI work, however “thin” and distinct from the ordinary categories (eg Music, Literature etc) it may be. Indeed I would question how much the judges involved actually understand this enough.
Does anyone know how this policy is tracking with or affected by policies in whether the AI engines themselves are infringing copyright?
While my thinking is in line with what @[email protected] and @[email protected] have already said, why can't "AI artists" just do what everybody in a profit-seeking situation does and just lie about it? "No your honor, our studies have shown cigarette smoking is not hazardous to your health," "yes, your honor, OxyContin is completely safe," or in this case "yes, your honor, I created this illustration." If your conscience is really bothering you, you could claim it was AI-assisted. I wouldn't think there'd be a "Big Eyes" prove-you-painted-that courtroom case. Am I wrong?
It detectable by actual artists, trained eyes, and probably other ai algos. Prompters have no true artistic skill 80 percent of the time, so they rely on AI to even do the finer details.
There are a few actual artists who use AI in the way scammers do (rather than as a tool to enhance their workflow), but they are rare.
Our thinking is not that different. There is a world of difference of describing an image and creating it.
But I have to strongly disagree with the rest of your assessment. Beyond the flood of six-fingered waifu, I've seen some beautiful, legitimate works created by this new tool. If you have to use an "ai algo" you've already defeated the purpose. It's illustrations we're discussing here, not banknotes.
BTW, do you consider Photoshop/Krita/GIMP artists "scammers"? Blender/Maya/Cinema4D artists? Who are these "actual artists" of which you speak?
In any case, we're still in uncharted territory. And personally I'm not crazy about the work in question. It LOOKS (by my "actual artist trained eyes") AI generated, regardless of the human Photoshop retouching involved.
Then auto-tuned, sample-ridden, garbage "music" should follow the same rules. The "artist" that "made" those songs shouldn't have any rights to the material as they didn't make anything. They had a computer rework their voice, who's words are often someone else's, with music that was also made by someone else.
Someone made the reference images generative AI used to make the end result image just like someone made the music that Lil' NoTalent69420 copied used for their song. Lil' NoTalent69420 manipulated it as much or even less than the individuals making AI art. It's pretty frustrating.