The hiccup is that the statement is that pure AI output is not subject to copyright, but once a human does something ‘transformative’ to it, then that is. So if a human can work back to the untouched AI output, they can go to town but they can’t reproduce any modification that was done after the AI generation. Since in practice, the GenAI row output is not provided it’s impossible for the human to know what the human touches were to undo, so that scenario does not present a whole lot of realistic implications.
The person who built it owns it, and can transfer that ownership to the person that paid for it.
If a machine made it, no one owns it.
Copyright applies to creative works, which are defined as a manifestation of creative effort in the world through a creative process involving one or more individuals. It can be argued that since no human creativity was involved, copyright does not apply.
so, everything’s fair game since the industrial revolution?
look, I understand what you’re trying to say and for the most part I agree but my point is that AI is fundamentally just another tool doing the same thing as other tools but unified and using natural language. what’s the difference if someones builds a website using a website building app like Wix, or uses AI to build one?
i still fundamentally disagree with the enormous environmental costs of AI, the infringements of creatives and contributors in its training, the economic cliff its disruption has put us on, the rapid expansion of energy-hungry infrastructure data centres, and the health reports of local residents who have been forced to endure the 24-hour operation of them.
To answer your point, creating an object using a machine is not a creative process by the definition used in this context, unless you put your creativity in it too. It’s the difference between creating a dress based on your own design and making it based on someone else’s. The ownership comes from the ideation, not the machanical process used to make the dress.
The key difference between AI and other tools is that it replaces the creative part of the process, thus removing the human factor.
As a side note, most of the time the end product contains both elements generated with AI and others done by humans. How much human contribution is required for it to be considered a creative work will probably be up to courts and lawmakers I’m afraid.
On a more ideological note, one could argue that AI models were created using basically the entirety of humanity works, often without compensating the ones holding the rights to such works, and thus should be owned by all humanity.
“creativity” is not as easy to define as you seem to imply. When photography replaced a lot of portrait painters, does that mean it wasn’t creative work? I mean sure painters can still paint other things, but likewise, coders can still code without AI. Does this just mean that there is “non-creative” code (the code that AI can generate) and “creative” code?
I said in my reply that determining whether something that was made with both AI and human contributions should be deemed a work of human creativity is not trivial. That being said, I can’t consider a piece of code entirely written by AI to fit the criteria for copyright protection.
And, dude, photography is 100% creative work and recognized as such everywhere. Yes, the machine takes the picture, but there’s a human who decided the subject, how to frame it, light, composition, not to mention the artistic intent, what he means to convey with that picture.
Then is AI code creative because the human wrote the prompt or guided the agent? But as you already said you agree that creativity is hard to define, I’m ok with us disagreeing on the exact boundary.
criteria for copyright protection.
honestly copyright was already BS as is. “Fair use” is such a mess
Then is AI code creative because the human wrote the prompt or guided the agent?
That’s something that is and will be argued about to the death. I don’t think that a prompt is enough to say the person writing it had a part on the creation of the thing, but I know someone will disagree. I see it as someone commissioning a portrait from a painter. No matter how detailed the request is, no one would say that the customer painted the picture.
Aside from that, I guess what I’m trying to say is that, in my view, creativity is something inherently human, and so can’t be attributed to a machine or software by definition.
So what you’re saying is that there’s a tipping point in “intellectual property” where after a certain amount of creative help from a machine it stops being your idea/property?
My mind goes to real play DND podcasts. There is a framework that allows them to tell their story, but the story still belongs to them.
So like if an LLM writes a function to make the program work properly, like to avoid a logic error you didn’t forsee without you asking it too. You don’t own that function? Would it be different if you asked it to fix that bug, or do you have to say how to fix the bug?
Well, probably? I’m by no means an expert, so I don’t know how that could be defined or work in practice. And I think it’ll take us a long time to figure this one out
This has also come up for generative AI with consistent result.
Problem in those cases is that they hinged upon the human passing through the non-human product without even claiming to have modified it. If human modification has been done, then that modification is subject to copyright. In the case of a composite of human and non-human origination, good luck extracting the non-human portion of the work to take advantage of that copyright opening.
I agree it is a crap comparison, just warning that even as the slop straight from the LLM is not copyrightable, blend in a little human touch up and it is “blessed” into being copyrighted. So it’s of little consequence unless the operator just passes it through untouched or makes it somehow easy to identify and remove the human tweaks.
In terms of slopping up a rip off under the assumption that it would obviously be the same as the original slop operators starting point, AI companies in their defense have already stated that operators can violate copyright implicitly as they legally declared they aren’t on the hook for any resultant infringement…
I wonder if their legal “protection” will work in court.
Intentionally absurd comparison: if I would host all the batman movies on a public website but make people agree to they are legally responsible for illegally downloading the movies, would I get away with that? I don’t think so.
Consequently, how much distance does there need to be between illegal activity and the company/person facilitate that activity?
Commenter is right. It remains to be seen whether or not it’s actually possible to license code generated this way as anything other than like MIT.
A person didn’t really make it, and so a person doesn’t really own it.
All AI generated code should be GPL. It was trained on all of us, and its output belongs to all of us.
If the model was trained on GPL code then one could argue anything the model outputs is a derivative.
That would be a copyright violation.
You can’t take closed-source code, make a derivative of it, and then GPL it. That’s illegal
MIT code can be used in GPL projects though. It simply becomes GPL code.
The hiccup is that the statement is that pure AI output is not subject to copyright, but once a human does something ‘transformative’ to it, then that is. So if a human can work back to the untouched AI output, they can go to town but they can’t reproduce any modification that was done after the AI generation. Since in practice, the GenAI row output is not provided it’s impossible for the human to know what the human touches were to undo, so that scenario does not present a whole lot of realistic implications.
a person pays another person to build them a website. who owns it?
edit: [everyone disliked that]
The person who built it owns it, and can transfer that ownership to the person that paid for it.
If a machine made it, no one owns it.
Copyright applies to creative works, which are defined as a manifestation of creative effort in the world through a creative process involving one or more individuals. It can be argued that since no human creativity was involved, copyright does not apply.
so, everything’s fair game since the industrial revolution?
look, I understand what you’re trying to say and for the most part I agree but my point is that AI is fundamentally just another tool doing the same thing as other tools but unified and using natural language. what’s the difference if someones builds a website using a website building app like Wix, or uses AI to build one?
i still fundamentally disagree with the enormous environmental costs of AI, the infringements of creatives and contributors in its training, the economic cliff its disruption has put us on, the rapid expansion of energy-hungry infrastructure data centres, and the health reports of local residents who have been forced to endure the 24-hour operation of them.
I probably could have worded it better.
To answer your point, creating an object using a machine is not a creative process by the definition used in this context, unless you put your creativity in it too. It’s the difference between creating a dress based on your own design and making it based on someone else’s. The ownership comes from the ideation, not the machanical process used to make the dress.
The key difference between AI and other tools is that it replaces the creative part of the process, thus removing the human factor.
As a side note, most of the time the end product contains both elements generated with AI and others done by humans. How much human contribution is required for it to be considered a creative work will probably be up to courts and lawmakers I’m afraid.
On a more ideological note, one could argue that AI models were created using basically the entirety of humanity works, often without compensating the ones holding the rights to such works, and thus should be owned by all humanity.
“creativity” is not as easy to define as you seem to imply. When photography replaced a lot of portrait painters, does that mean it wasn’t creative work? I mean sure painters can still paint other things, but likewise, coders can still code without AI. Does this just mean that there is “non-creative” code (the code that AI can generate) and “creative” code?
I said in my reply that determining whether something that was made with both AI and human contributions should be deemed a work of human creativity is not trivial. That being said, I can’t consider a piece of code entirely written by AI to fit the criteria for copyright protection.
And, dude, photography is 100% creative work and recognized as such everywhere. Yes, the machine takes the picture, but there’s a human who decided the subject, how to frame it, light, composition, not to mention the artistic intent, what he means to convey with that picture.
Then is AI code creative because the human wrote the prompt or guided the agent? But as you already said you agree that creativity is hard to define, I’m ok with us disagreeing on the exact boundary.
honestly copyright was already BS as is. “Fair use” is such a mess
Am I a chef because I ordered a pizza and modified some toppings?
Haha I totally agree that copyright is a mess!
That’s something that is and will be argued about to the death. I don’t think that a prompt is enough to say the person writing it had a part on the creation of the thing, but I know someone will disagree. I see it as someone commissioning a portrait from a painter. No matter how detailed the request is, no one would say that the customer painted the picture.
Aside from that, I guess what I’m trying to say is that, in my view, creativity is something inherently human, and so can’t be attributed to a machine or software by definition.
So what you’re saying is that there’s a tipping point in “intellectual property” where after a certain amount of creative help from a machine it stops being your idea/property?
My mind goes to real play DND podcasts. There is a framework that allows them to tell their story, but the story still belongs to them.
So like if an LLM writes a function to make the program work properly, like to avoid a logic error you didn’t forsee without you asking it too. You don’t own that function? Would it be different if you asked it to fix that bug, or do you have to say how to fix the bug?
Well, probably? I’m by no means an expert, so I don’t know how that could be defined or work in practice. And I think it’ll take us a long time to figure this one out
Sounds like prompt logs need to be tracked in git with reference. But also it sounds basically unenforceable.
In the usa, it has been ruled in the past that only human can have copyright. And if a human is not the author, they don’t have the copyright.
https://en.wikipedia.org/wiki/Monkey_selfie_copyright_dispute
This has also come up for generative AI with consistent result.
Problem in those cases is that they hinged upon the human passing through the non-human product without even claiming to have modified it. If human modification has been done, then that modification is subject to copyright. In the case of a composite of human and non-human origination, good luck extracting the non-human portion of the work to take advantage of that copyright opening.
Yes. My point was that the person’s comparison is shit. That genAI’s and human’s output aren’t the same.
I agree it is a crap comparison, just warning that even as the slop straight from the LLM is not copyrightable, blend in a little human touch up and it is “blessed” into being copyrighted. So it’s of little consequence unless the operator just passes it through untouched or makes it somehow easy to identify and remove the human tweaks.
In terms of slopping up a rip off under the assumption that it would obviously be the same as the original slop operators starting point, AI companies in their defense have already stated that operators can violate copyright implicitly as they legally declared they aren’t on the hook for any resultant infringement…
I wonder if their legal “protection” will work in court.
Intentionally absurd comparison: if I would host all the batman movies on a public website but make people agree to they are legally responsible for illegally downloading the movies, would I get away with that? I don’t think so.
Consequently, how much distance does there need to be between illegal activity and the company/person facilitate that activity?