- cross-posted to:
- technology@lemmit.online
- cross-posted to:
- technology@lemmit.online
We Asked A.I. to Create the Joker. It Generated a Copyrighted Image.::Artists and researchers are exposing copyrighted material hidden within A.I. tools, raising fresh legal questions.
It has relevance to what counts as an original artwork.
This is what you said:
No it is not. They do not have enough differences to be considered original in any court of law.
???
If I ask for an image of Joaquin Phoenix as The Joker from the movie The Joker, then yes it will not be original.
If I ask for original drawings off original ideas it will be original.
Therefore AI can be used for both.
Therefore the technology itself is not infringing, but only specific uses of it are, same as with a VCR, an HDD and our very brains. This should be obvious, and NYT knows it’s going to lose and that’s why they are now developing their own model. This case is just to stall the industry until old money corpos can catch up to avoid being disrupted out of existence. It has zero legal ground
Again, VCRs and hard drives can’t create content. They can only capture content. AI can create content, but it is not always original. Which is the problem. No one is trying to sue them over things that are credibly original.
It is no more legal for you to tell an AI to make you a picture of the Joker as it is to ask a human artist to do it. And if the human artist did it, WB/DC would be within their rights to take them to court because it would violate both trademark and copyright. They usually don’t, but they are within their rights.
You can ask a VCR or a hard drive to draw you a picture of The Joker all day. They won’t because they can’t.
If AI was only capable of creating original artworks, this would not be an issue.
There is no difference, a camcorder creates content, but it doesn’t mean they’re banned just because you can film a film with one.
Yes I agree, it is not legal to replicate copyrighted works regardless.
But again, human artists aren’t illegal just because they can infringe copyright in a hypothetical scenario. Same with AI. The machine is non-infringing, the prompt operator can infringe copyright if they try, and then they are responsible under law, same as a human artist would be.
The machine is blameless regardless of what it was trained on.
Nope. Camcorders do not create content. They record content. Camcorders do not create anything. That is a ridiculous claim. I cannot point a camcorder at you and have it make you look like Heath Ledger.
AI creates content. It can make things that literally don’t exist. If I tell it to make me Heath Ledger as The Joker fighting Jack Nicholson as the Joker, it can create it. A camcorder can’t. A VCR can’t. A hard drive can’t. I have no idea why you don’t understand the difference between creating content and recording content.
I also said nothing about the AI itself being illegal, so I also have no idea where you’re getting that from. I said it is violating copyright and trademark when it creates such images. Because it is.
Hence the lawsuits. Hence the lack of such lawsuits against camcorders, VCRs and hard drives.
It flat out isn’t lol, I don’t know what to tell you bud, but you have no idea what you’re talking about, and there was a lawsuit against VCRs:
https://en.m.wikipedia.org/wiki/Sony_Corp._of_America_v._Universal_City_Studios,_Inc.
In fact it’s quite likely that the OpenAI decision will be based upon this.
It’s not that complicated to understand:
AI trains on images (fair use) -> Prompter inserts prompt -> output can be copyright infringing or not, depending on the prompt, same as a brain, hard drive, VCR and an HDD.
The metaphysics of what counts as creation vs recording are irrelevant, everything you’re saying is just flat out irrelevant.
Yes, there was a lawsuit against VCRs. It had nothing to do with original content.
Nope, that is not in any legal definition of fair use.
What prompt do I enter to get a VCR to make me a picture of Jack Nicholson’s Joker fighting Heath Ledger’s Joker?
Do I press both rewind and fast forward at once to access the secret content generation menu?
You dress up and film it with your buddies and record it to a second gen tape lol? Again, the content generation aspect is irrelevant, what matters is whether a piece of equipment is made to infringe copyright or not, AI isn’t, neither are VCRs, that’s law, simple as really.
And yes training is fair use, it better be, I train my brain on images all the time.
Which means it is not creating content. It is recording content. Which was my point.
Please back this up. Your brain is not a computer. Furthermore, even if it was, someone else would be training it and you cannot legally train someone else on copyrighted material that you have not licensed, which is why schools have to license textbooks and a teacher that teaches from an unlicensed textbook can be sued. That’s the entire impetus for the Open Textbook Library. The Open Textbook Library would literally not need to exist if training material was not protected by copyright.
You see, the problem here is that you keep claiming things that are the opposite of what these companies are getting sued for doing. And yet those suits aren’t getting laughed out of court. Doesn’t that tell you that maybe your ideas of how the law works here are wrong?
I have been studying U.S. copyright and trademark law for over 15 years. How long have you been studying it?