I'm pretty sure that that's covered under the option of "the machine itself", as that is a computer running an instance of the algorithm.
I'm actually a bit surprised that this legal territory is so poorly explored, I would have assumed that at least there would have been challenges, lawsuits, and court cases on the validity (and ownership) of copyright for procedurally generated artwork long before now (of course, this would still leave as an open question the issue of whether such a work infringes on works an AI was trained on).
and in 30 years Disney AI will be assimilating competing ones in hostile takevers. Yes your Honor, BeepBoop2060 is one with squeek12 who owned those works previously.
Who owns the product of an arthouse? Of a manufacturing plant? Of a farm?
The creators and/or owner of the algorithm own the product.
If they use it to train with other artists creation, that is the same as taking inspiration. If they release the alg open source, then its the same as any product that makes things, the maintainer of that instance owns the production of anything created from it.
That's not the point. Copyright law - at least the European versions of it - requires a human creator. So the output of an algorithm is not protected under copyright law in general, even if the output appears to be "creative". If the output includes parts of a pre-existing work by a human creator, then it might be subject to copyright law, but in favour of the creator of the pre-existing work.
Of course the output of an algorithm is covered, if there is creative input. Otherwise nothing generated from photoshop would be protected.
And creativity is the point. As a previous post mentioned, "creative powers of the mind"
Collage may be a better analogy than paint. An artist can make an image with a collage of bits of other artists' works, in much the same way that a generative network can make an image out of image elements derived from other artists' works.
The question is not whether the "AI" has any ownership, but whether the original artists do, or whether the programmers do, or whether the people selecting training images do, or whether the person selecting a final artwork out of a mass of inferior generated images does.
> An artist can make an image with a collage of bits of other artists' works, in much the same way that a generative network can make an image out of image elements derived from other artists' works.
That's not a good analogy because a human is involved. The entire point here is whether living entities other than persons can have copyright, and the answer (in the US) is no.
"living entities other than persons" having or not having copyright is unrelated, as programs are not living entities. Humans are definitely involved in developing, training, operating, and choosing output from machine learning art-generating systems.
What you say makes sense in principle, but I don't think it sufficiently accounts for the fact that with AI we can mass generate enormous quantity of artwork with essentially zero incremental effort.
If I create an AI that creates billions of different image of dogs sitting on a park bench, and then someone else creates an image of a dog sitting on a park bench, odds are pretty high that it would look substantially similar to at least one of my AI generated images. Would it then be fair for me sue whoever is using that image for copyright infringement?
You certainly can sue, but you would be expected to fail because accidentally looking substantially similar is not copyright infringement.
There is some precedent for substantial similarity in certain popular songs, but it is based on convincing the court of the argument that the author of the later work must have had heard the original one (for example, over the radio) and copied parts of it, possibly unintentionally.
However, if it's reasonable to demonstrate that in your proposed case that "someone else" had not seen that particular one-in-a-billion image of dogs sitting on a park bench (since probably noone had ever looked at that particular image out of the billions you generated), then their creation is not a copy of your image ("bits have color" - even if it would happen to be perfectly identical, it's not a copy if it was not actually copied) nor derived work of your image and so it does not infringe upon your copyrights no matter how similar it looks.
Running a program which generates art is no different from clicking the shutter on a camera, or throwing paint on a canvas randomly (non-repeatable physical process generates the image.
Yeah people get too caught up in the marketing-hype of AI as if it's a magical thinking lifeform. So far, it is not. It's a device, a tool. Devices don't own things.
The point is not that devices don't own things but that the owner of the device / creator of the software can't copyright its outputs in his name.
Generating art costs money - electricity to train models and rent/buy the hardware, requires expertise in neural nets, tons of training data from the public domain or privately created, lots of experimenting and fiddling with the parameters, manual curation of the output ...
why should the human who invested all this work not be entitled to the copyright?
Whoever wants something very specific that is still hard to reproduce/generate.
In the photography industry there are billions of free images all over the internet, who's going to pay for the next image? Only those who want something specialised that can't be had for free.
Copyright laws of most European countries specify that only works created by human beings can be afforded copyright protection. The same seems to apply to current US laws (which I'm not familiar with in detail). In UK and Ireland it is possible in principle that computer generated works can be copyrighted by the person responsible for the creation of the work. So it depends on the country. Notwithstanding the above, a work may of course be sold, in this case by the owner of the printed sheet of paper.
The U.S. Copyright Office has already declared that if it's generated by a machine, without any creative input or intervention from a human author, then the work is NOT eligible for copyright at all.
They believe that "The copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind.” Trade-Mark Cases, 100 U.S. 82, 94 (1879). Because copyright law is limited to “original intellectual conceptions of the author,” the Office will refuse to register a claim if it determines that a human being did not create the work. Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 58 (1884). ... the Copyright Act protects “original works of authorship.” 17 U.S.C. § 102(a) (emphasis added). To qualify as a work of “authorship” a work must be created by a human being. See Burrow-Giles Lithographic Co., 111 U.S. at 58. Works that do not satisfy this requirement are not copyrightable. The Office will not register works produced by nature, animals, or plants.... Similarly, the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author."
A future US court ruling or law change could change this, of course, but I sure wouldn't bet on it. The US Copyright Office has lawyers who specialize in this area, have clearly stated their position, and they are citing specific laws and court cases to justify their position. Also, in most cases you can't go forward with a copyright claim without registration (though you can register after-the-fact); if the US Copyright Office won't allow registration, you'll have a really challenging time even starting a court case.
I'm sure there are edge cases and complications. But the essence is only "unsettled" if you don't look for the existing rulings. A Google search will find actual answers from credible, authoritative sources. Sure, courts and Congress could change it, but it'd require a significant change.
While this information is specific to the U.S., copyright is similar enough in other jurisdictions that there's a decent chance that other jurisdictions would come to the same conclusion. European law in particular has a number of similarities in this area; I'd love to hear from those with more background in other jurisdictions.
Disclaimers: I am not a lawyer, and this is not legal advice.
I don't know if hyperparameters count as being adequately creative, and perhaps that is unsettled. I suspect you would have to have an argument about how creative the process of tuning parameters is. It might be very fact-specific, IE perhaps some hyperparameters might count as creative and others might not. I realize that is not very satisfying.
The US Copyright Office says: "The crucial question is “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.” U.S. Copyright office, Report to the Librarian of Congress by the Register of Copyrights (1966)."
While the dividing line is fuzzy, it is clear that as soon as you're claiming that the art was generated by AI, you are fighting an uphill battle to get a copyright registered.
I think it is perhaps interesting to compare to photography. There a human picks the initial parameters (position, angle, lens, time of day), presses a button, and then the camera (a device) does all the work. It is not controversial that photographers have a copyright on their work.
You're right, it is not controversial today that photographers have a copyright on their work. Which is why the monkey selfie is so important. In the monkey selfie case, a photographer did set up the equipment, which means that some of the parameters were indeed established by a human. However, the monkey actually created the picture, including the timing and body placement. The copyright office has determined that the work could not be copyrighted.
Which leads to the conclusion that at least in some cases,a human doing some setup is not enough to grant a copyright.
While this case was partly about the question "can a Monkey own a copyright in a selfie it creates?", I think it got a lot of press because it was so obviously related to the case of AI-created art. If one non-human can get a copyright, why not another? The US Copyright office has shot them both down, on the grounds that only humans can get copyrights.
In the monkey selfie case, the photographer set up the camera, which meant that there were some very abstract settings that involved a human. That wasn't enough to get a copyright. I think at least some hyperparameters would be similar - not really enough to justify creativity.
Within the EU, this is not the case - and quite the opposite situation.
The EU copyright directive from 2003 rules that only a natural person (a human, not a company) can have copyright ownership on its works. Companies only can own perpetual licenses for these works.
Generated art where a machine writes code or produces artwork based on randomization leads to the inventor of the algorithm owning the copyright for it.
As there is nothing like "public domain" in the EU area possible, you can only distance yourself from this work with the CC0 license while stating that you cannot offer any warranty and that you cannot be held responsible for malicious use of the product and that users of the product use it with their own risk.
Source: wrote a code generating engine based on backpropagated ES/HyperNEAT, got sued in front of the EU court because it turned out it got quite good at pentesting, too.
I was curious and read that linked wikipedia article on the monkey selfie, which actually makes it clear that US law pretty much does have an answer to the question whether an animal can be a copyright owner.
The actually interesting part is the fact that these photos/selfie situation were set up specifically by a photographer with their equipment.
That is actually the relevant (to AI) portion in all this. If someone sets up all the right pieces but then has a subject or generally "something" else cause the actually creation of the artwork the lines get really blurry. Hypothetically if you set up an elaborate set of machinery guiding an animal/machine/human to paint a picture it wouldn't otherwise be painting, e.g. think elephant with brushes/paint on trunk walks through a guided maze of a huge canvas, the coordinator would not have copyright (by the logic of the monkey case).
P.S. While I sympathize with the photographer, the level of butthurt and entitlement is just rich:
> Slater was unable to travel to the July 2017 court hearing in the United States for lack of funds and said he was considering alternative careers as a dog walker or tennis coach.[52] "This would be a new venture for me. It would pay peanuts, but at least it would be more than photography. I am just not motivated to go out and take photos any more. I've had outlays of several thousand pounds for lawyers, it is losing me income and getting me so depressed. When I think about the whole situation I really don't think it's worth it."[52] Slater added, "Everything I did to try and highlight the plight of the monkeys has backfired on my private life. I've had my life ruined."[52]
They made an initial investment, made good on it, then decided to sue Wikimedia without the necessary funds. This is a very special case and not at all related to the viability of the wildlife photographer profession. Nothing about this means the photographer has to abandon their career and start walking dogs instead.
The list is nonsensical. If there is a copyright, it is owned by whoever is running the software. The legal owners. Any program with emergent properties could technically produce output that could be construed as art. Such programs have been around since decades. It isn't like we are on the verge of some new era where we have to consider computers or algorithms as equals or peers that deserve respect.
I suspect this mentality will not age well. We very well may be close to discussing robot rights, and the answer to this question in 50 years might be simple: the machine owns the copyright.
I believe this could depend on the complexity of AI. If we're talking about 'real' artificial intelligence that is self-aware and can mimic human thinking, can interact with human beings as we do with one another then perhaps at first we should recognize AI as natural person or entity, or legal person and grant them necessary rights; which later would lead to all sort of legal issues including copyrights. But if we're talking about AI that is just an algorithm that automates tasks then perhaps its creators should hold the rights or the work should become part of public domain.
47 comments
[ 4.0 ms ] story [ 84.2 ms ] threadI'm actually a bit surprised that this legal territory is so poorly explored, I would have assumed that at least there would have been challenges, lawsuits, and court cases on the validity (and ownership) of copyright for procedurally generated artwork long before now (of course, this would still leave as an open question the issue of whether such a work infringes on works an AI was trained on).
The creators and/or owner of the algorithm own the product.
If they use it to train with other artists creation, that is the same as taking inspiration. If they release the alg open source, then its the same as any product that makes things, the maintainer of that instance owns the production of anything created from it.
https://web.archive.org/web/20200613105427/https://aimade.ar...
Many works of art involve processes and the line between "process" and "generated" is very fuzzy.
The question is moot. The tool cannot own the copyright.
The AI is no different than a brush or a hammer or a screwdriver.
The question is not whether the "AI" has any ownership, but whether the original artists do, or whether the programmers do, or whether the people selecting training images do, or whether the person selecting a final artwork out of a mass of inferior generated images does.
That's not a good analogy because a human is involved. The entire point here is whether living entities other than persons can have copyright, and the answer (in the US) is no.
If I create an AI that creates billions of different image of dogs sitting on a park bench, and then someone else creates an image of a dog sitting on a park bench, odds are pretty high that it would look substantially similar to at least one of my AI generated images. Would it then be fair for me sue whoever is using that image for copyright infringement?
There is some precedent for substantial similarity in certain popular songs, but it is based on convincing the court of the argument that the author of the later work must have had heard the original one (for example, over the radio) and copied parts of it, possibly unintentionally.
However, if it's reasonable to demonstrate that in your proposed case that "someone else" had not seen that particular one-in-a-billion image of dogs sitting on a park bench (since probably noone had ever looked at that particular image out of the billions you generated), then their creation is not a copy of your image ("bits have color" - even if it would happen to be perfectly identical, it's not a copy if it was not actually copied) nor derived work of your image and so it does not infringe upon your copyrights no matter how similar it looks.
Generating art costs money - electricity to train models and rent/buy the hardware, requires expertise in neural nets, tons of training data from the public domain or privately created, lots of experimenting and fiddling with the parameters, manual curation of the output ...
why should the human who invested all this work not be entitled to the copyright?
In the photography industry there are billions of free images all over the internet, who's going to pay for the next image? Only those who want something specialised that can't be had for free.
The U.S. Copyright Office has already declared that if it's generated by a machine, without any creative input or intervention from a human author, then the work is NOT eligible for copyright at all.
They believe that "The copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind.” Trade-Mark Cases, 100 U.S. 82, 94 (1879). Because copyright law is limited to “original intellectual conceptions of the author,” the Office will refuse to register a claim if it determines that a human being did not create the work. Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 58 (1884). ... the Copyright Act protects “original works of authorship.” 17 U.S.C. § 102(a) (emphasis added). To qualify as a work of “authorship” a work must be created by a human being. See Burrow-Giles Lithographic Co., 111 U.S. at 58. Works that do not satisfy this requirement are not copyrightable. The Office will not register works produced by nature, animals, or plants.... Similarly, the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author."
I'm quoting chapter 300 here: <https://www.copyright.gov/comp3/chap300/ch300-copyrightable-... of the Compendium of U.S. Copyright Office Practices here <https://www.copyright.gov/comp3/>. The compendium is the "governing administrative manual for registrations and recordations issued by the U.S. Copyright Office." In short, for the US, this is an extremely authoritative source.
A future US court ruling or law change could change this, of course, but I sure wouldn't bet on it. The US Copyright Office has lawyers who specialize in this area, have clearly stated their position, and they are citing specific laws and court cases to justify their position. Also, in most cases you can't go forward with a copyright claim without registration (though you can register after-the-fact); if the US Copyright Office won't allow registration, you'll have a really challenging time even starting a court case.
I'm sure there are edge cases and complications. But the essence is only "unsettled" if you don't look for the existing rulings. A Google search will find actual answers from credible, authoritative sources. Sure, courts and Congress could change it, but it'd require a significant change.
While this information is specific to the U.S., copyright is similar enough in other jurisdictions that there's a decent chance that other jurisdictions would come to the same conclusion. European law in particular has a number of similarities in this area; I'd love to hear from those with more background in other jurisdictions.
Disclaimers: I am not a lawyer, and this is not legal advice.
The US Copyright Office says: "The crucial question is “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.” U.S. Copyright office, Report to the Librarian of Congress by the Register of Copyrights (1966)."
While the dividing line is fuzzy, it is clear that as soon as you're claiming that the art was generated by AI, you are fighting an uphill battle to get a copyright registered.
Again, I am not a lawyer.
Which leads to the conclusion that at least in some cases,a human doing some setup is not enough to grant a copyright.
While this case was partly about the question "can a Monkey own a copyright in a selfie it creates?", I think it got a lot of press because it was so obviously related to the case of AI-created art. If one non-human can get a copyright, why not another? The US Copyright office has shot them both down, on the grounds that only humans can get copyrights.
In the monkey selfie case, the photographer set up the camera, which meant that there were some very abstract settings that involved a human. That wasn't enough to get a copyright. I think at least some hyperparameters would be similar - not really enough to justify creativity.
The EU copyright directive from 2003 rules that only a natural person (a human, not a company) can have copyright ownership on its works. Companies only can own perpetual licenses for these works.
Generated art where a machine writes code or produces artwork based on randomization leads to the inventor of the algorithm owning the copyright for it.
As there is nothing like "public domain" in the EU area possible, you can only distance yourself from this work with the CC0 license while stating that you cannot offer any warranty and that you cannot be held responsible for malicious use of the product and that users of the product use it with their own risk.
Source: wrote a code generating engine based on backpropagated ES/HyperNEAT, got sued in front of the EU court because it turned out it got quite good at pentesting, too.
I could offer to talk about it via telegram secure chat, if you're interested. I have the same pseudonym there.
The actually interesting part is the fact that these photos/selfie situation were set up specifically by a photographer with their equipment.
That is actually the relevant (to AI) portion in all this. If someone sets up all the right pieces but then has a subject or generally "something" else cause the actually creation of the artwork the lines get really blurry. Hypothetically if you set up an elaborate set of machinery guiding an animal/machine/human to paint a picture it wouldn't otherwise be painting, e.g. think elephant with brushes/paint on trunk walks through a guided maze of a huge canvas, the coordinator would not have copyright (by the logic of the monkey case).
P.S. While I sympathize with the photographer, the level of butthurt and entitlement is just rich:
> Slater was unable to travel to the July 2017 court hearing in the United States for lack of funds and said he was considering alternative careers as a dog walker or tennis coach.[52] "This would be a new venture for me. It would pay peanuts, but at least it would be more than photography. I am just not motivated to go out and take photos any more. I've had outlays of several thousand pounds for lawyers, it is losing me income and getting me so depressed. When I think about the whole situation I really don't think it's worth it."[52] Slater added, "Everything I did to try and highlight the plight of the monkeys has backfired on my private life. I've had my life ruined."[52]
They made an initial investment, made good on it, then decided to sue Wikimedia without the necessary funds. This is a very special case and not at all related to the viability of the wildlife photographer profession. Nothing about this means the photographer has to abandon their career and start walking dogs instead.
A rights!