An artist who infamously duped an art contest with an AI image is suing the U.S. Copyright Office over its refusal to register the image’s copyright.
In the lawsuit, Jason M. Allen asks a Colorado federal court to reverse the Copyright Office’s decision on his artwork Theatre D’opera Spatialbecause it was an expression of his creativity.
Reuters says the Copyright Office refused to comment on the case while Allen in a statement complains that the office’s decision “put me in a terrible position, with no recourse against others who are blatantly and repeatedly stealing my work.”
Why do you say your employer has no intent? They hired you with a particular product in mind after all. And they can do everything that you do with an AI: evaluate the results and tell you to try again.
What, specifically, do you do with an AI that an employer cannot do with a photographer?
It depends on who is creating the art. If I simply go where I am told, take exactly the pictures that I am told to take, and then hand them over to the employer, I am little more than a tripod that can work a shutter. I’m not the one with the intent, my employer is, and in that case I would be the tool (and my exes agree). In that case, my employer has effectively made the art, I have had no input in the process, and am for all intents and purposes no different than the AI. However, when I start to force my will upon the photographs, when I stage the lighting and set everything the way that I wish, then I am the one with intent, and I have created the art, and the camera is the only tool in the bag. That is how the art shifts; with the intent. Every pencil’s lead will have a grain specific to it, like a fingerprint, however, that pencil has no control over the art that it creates. It is a tool. If the AI, the pencil, or the photographer start to exert their own will over the art, then it becomes their own. If they take suggestion from someone else, like an employer, then a contract and shared creative ownership can be argued, but that is something between two sentient life forms. A tool cannot own the art. I do not credit my camera for the shots I take.
This sounds like a very easy test for an employer to pass. They force their will simply by telling you what to shoot.
But I gather that you won’t give them ownership quite so easily, they need to control every aspect of how you take the photos and thus reduce you to a “tripod”.
You can’t have two standards. Which is it? If merely exerting will is enough, then employers always own what photographers produce. If some degree of independence beyond a tripod allows the photographer to claim ownership, then AI users can’t claim ownership.
Can you articulate a single principle that is valid for both employers and AI users?
I actually covered this exact point:
When two sentient life forms collaborate to create art, then they share the ownership. When an employer tells me what photographs to take, they have a part in the creative process and have placed their intent into the work. Now, 99% of the time, when an employer asks for something, and I do the work, they don’t take the credit for it. They defer to me and understand that my knowledge has given their idea form, and because of my intent, their intent has been realised. However, there can be arguments made for shared ownership if they have given me input as to what they want. I’ll even praise someone who has done some research beforehand and said something like, “When I was here last week, at 4pm, the sun shone through here beautifully, and we’d like to get some photos with that.” That is a shared creative experience, and the intent of both parties creates the art, and so both parties have some ownership. You can also look at this through the lens of the music industry, where a performer may not write their own songs, and both the artist and the songwriter share credit for the song (though usually not equally).
Now, when I give instructions to an AI on what art I would like to see, that AI has no input in the process, it simply pulls from its dataset and applies a randomly generated seed to create the image. It exerts no will of its own, and so no intent of its own is wielded over the art, as it has neither. It is no more willful than the grain of the pencil lead.
I think your approach would not work in practice. The test is not how it plays out when people are cooperating, but what happens when there is a dispute. And if the principle is “providing some input gives ownership” then the photographer, photographer’s assistant, agent, employer, and employer’s ex-wife will all sue each other over ownership.
In the music industry, you need to actually perform a piece to claim performance credit or specify the verses of a song that you personally wrote to claim writing credit.
I mean, that’s any artistic industry, really. Movies aren’t solely made by the director, music isn’t solely made by the singer. Sometimes those people can be the sole creator of the art, but when they aren’t, credit is shared.
Agreed, in which case you would get performance credit and everyone else will get credit for what their contributed. No one gives credit to the microphone cord, though. No one is crediting the studio lights. They aren’t sentient. Their intent isn’t exerted over the art.
Movies aren’t made solely by the director, but certain requirements must be met before one can claim copyright. Hundreds of people can offer their input but not be eligible for copyright, because offering input is not sufficient. There must be some direct control over an element of the output, whether that’s the cinematography, writing, or soundtrack.
It’s true that inanimate objects can’t claim copyright but that does not remove the requirement for direct control. If no human has direct control then the rights revert to public domain, for example no human has direct control of a sunset so a sunset cannot be copyrighted.
Right, and hundreds of people may have programmed the LLM, but they don’t get credit for the art.
A human does have direct control, though. I control the keywords. I control the random seed if I don’t want it to be random. In the case of MidJourney, I can prompt with an image to control the character, style, and over-all image composition. I have a lot of control over what comes out. Just because I don’t control exactly where each pixel goes doesn’t mean my intent isn’t exerted over the final piece, just like I can’t control every bristle on a paintbrush.
You directly control every pixel on your paintbrush, whether you want to or not. Who else controls it? It can only move when your mouse moves, which can only move when you cause your hand to move.
In contrast, you have some control over MidJourney output, but not direct control. Something could appear in the output that you did not cause.