Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
The Courts DRM

AI-Generated Works Aren't Protected By Copyrights, US Judge Rules (billboard.com) 28

A U.S. federal judge "ruled Friday that U.S. copyright law does not cover creative works created by artificial intelligence," reports Billboard magazine: In a 15-page written opinion, Judge Beryl Howell upheld a decision by the U.S. Copyright Office to deny a copyright registration to computer scientist Stephen Thaler for an image created solely by an AI model. The judge cited decades of legal precedent that such protection is only afforded to works created by humans. "The act of human creation — and how to best encourage human individuals to engage in that creation, and thereby promote science and the useful arts — was ... central to American copyright from its very inception," the judge wrote. "Non-human actors need no incentivization with the promise of exclusive rights under United States law, and copyright was therefore not designed to reach them."

In a statement Friday, Thaler's attorney Ryan Abbot said he and his client "disagree with the district court's judgment" and vowed to appeal: "In our view, copyright law is clear that the public is the main beneficiary of the law and this is best achieved by promoting the generation and dissemination of new works, regardless of how they are created."

Though novel, the decision was not entirely surprising. Federal courts have long strictly limited to content created by humans, rejecting it for works created by animals, by forces of nature, and even those claimed to have been authored by divine spirits, like religious texts.

The Hollywood Reporter notes that "various courts have reached the same conclusion." In another case, a federal appeals court said that a photo captured by a monkey can't be granted a copyright since animals don't qualify for protection, though the suit was decided on other grounds. Howell cited the ruling in her decision. "Plaintiff can point to no case in which a court has recognized copyright in a work originating with a non-human," the order, which granted summary judgment in favor of the copyright office, stated.
This discussion has been archived. No new comments can be posted.

AI-Generated Works Aren't Protected By Copyrights, US Judge Rules

Comments Filter:
  • by xack ( 5304745 ) on Saturday August 19, 2023 @02:42PM (#63780566)
    By "mining" copyright by ai generating randomly generated prompts and then copyrighting the result. By forcing human effort we can at least keep the delusion of copyright alive, despite its flaws.
  • For it to become copyright protected I'm assuming it would have required some human thought on the entire thing.
    Interestingly you could template away this problem as detection tools are looking for certain markers.
    Mind you, if you're writing the prompts, modifying 80% of the content or writing 80% of the content then you'll probably see this anyway.
    • by dbialac ( 320955 )
      And you have to be able to pass the copyright on through inheritance The AI is the author and doesn't have a family tree nor other entities to pass ownership on to. See also the monkey selfie.
    • Creative input and decision making is required. It doesn't need to be much, but it needs to exist. In the monkey copyright case it was obvious the monkey couldn't own the copyright, but while it wasn't litigated if the photographer would lose copyrights, the prevailing legal opinion was that simply the act of setting up a camera so it can create a certain exposure and then having the monkey steal it would be grounds for saying the photographer had enough creative input to qualify for a copyright. Now if the

  • So I guess those artists who use their fingers to draw paintings with their own blood are reasonably safe to be covered by "Copyright". All others will first need to prove their non-human tools are not sophisticated enough to be called "AI".

    I mean, US Copyright is a big mess anyway, and I am not unhappy if it goes away. But it is funny to suddenly see lines drawn between non-AI and AI tools - so was the "Camera Lucida" a legitimate painter's tool, or can you not claim a copyright if you used it to do a pai
    • by thegarbz ( 1787294 ) on Saturday August 19, 2023 @03:10PM (#63780636)

      But it is funny to suddenly see lines drawn between non-AI and AI tools - so was the "Camera Lucida" a legitimate painter's tool, or can you not claim a copyright if you used it to do a painting?

      No such line has been drawn. The whole claim from Stephen Thaler is that the image was entirely created by AI, i.e. without human input or prompt but rather a product of the AI's own sentience. There is no situation here where you can call the AI here a "tool" anymore than I can an architect and building company a "tool" for magicking a house they designed into existence.

      If you use Photoshop's AI fill tool you don't suddenly loose the ability to copyright. If Photoshop becomes self aware and starts drawing things without your prompt while you're sleeping then it fits in this scenario: no human creation.

      • To add to this, the actual functioning of the software and whether a human coded it is irrelevant. What is relevant is the claims which were made and that claim is the image is the product of sentient AI with no human prompts.

        • In that case it's simple: the copyright can be assigned only if the AI consents and gets some kind of consideration fo the assignment. If you want to argue the work was "for hire" then you'd have to show the AI has an employment agreement.

          The other option is that the creator is indeed the "operator" of the machine, and copyright should be assigned to the operator. Either that or the work is simply not copyrightable at all.

      • Further, their argument for not granting to the AI itself wasn't so much that the AI isn't human, but that the AI would not be motivated to work harder if it were given exclusive rights.

      • The whole claim from Stephen Thaler is that the image was entirely created by AI, i.e. without human input or prompt but rather a product of the AI's own sentience.

        Then the sentient AI should be able to explain to the court that it deserves protection.

        Thaler used a tool to create a something, and putt together this bullshit claim & lawsuit to generate publicity to earn money from. It's a scam.

        • I agree with you, but that isn't what this case is about. Thaler wasn't in court to argue his AI was sentient and thus didn't need to prove to the court that it was. He was there to argue about copyright and made the claim about the AI being sentient. That's the only thing the court has ruled on, and also my fundamental point. The OP claimed a line was drawn. No line has been drawn for tools. The court ruled on the facts presented to it based on widely accepted existing precedent, an AI is not human, real o

  • Write your own prompts folks.
  • Judge better watch their language - that's too specific. Every AI product so far is the result of a human making a request of a trained model.

    That request is a creative input that is NOT from the AI.

    • by dmay34 ( 6770232 )

      There is quite a lot of judicial case history regarding what does and does not constitute a copyright. This judge was not pulling this out of their butt. Copyright lawyers are not likely surprised by this. Just typing in a sentence into a prompt was never going to be enough creative effort to justify a copyright. Similarly to how you can't get a copyright on a line drawn across a page with a ruler, there is a certain level of creative effort that is needed to justify winning a copyright claim.

      • IANALawyer, I just occasionally scream at them (and judges) on the Internet.

      • This case had nothing to do with anyone inputting a prompt: it was specifically "promptless", as the AI was designed to create images with no human intervention. No prompt. THAT is why the judge ruled as they did.

    • by Luckyo ( 1726890 )

      Completely correct. However in this specific case, there was no such input.

      This is a very narrow edge case where the person defending the copyright specifically testified that there was no human input necessary, entire creation from start to finish was done without any human input.

      This is why you can copyright generative AI work that creates content based on your input.

    • Judge better watch their language - that's too specific.

      The law *is* specific and the judge's language was carefully chosen to be specifically relevant to this case. What AI is or isn't is absolutely irrelevant. It was the plaintiff's argument that the work was created solely by AI and the judge ruled based on this argument directly.

    • The AI code is already covered by copyright. And if they patented any of their inventions, then that would have been protected by the patent.

      This isn't about the AI itself. That is already protected.
    • by evanh ( 627108 )

      More importantly, the training material is copyrighted. So the AI company should be forking over a ton of money for all that source material.

  • Set AI's to generating everything and anything possible and then none of it can be copyrighted leaving it all open source and easily copyable by anyone. No more copyright because there will be too much free material. Maybe someone could write a proof that any particular text could have been generated by AI therefore it is not copywritable. That will lead to less need for judges and lawyers, so win-win. I think I'm being sarcastic but I'm not sure.
  • This will be interesting to watch, just to see how judges will eventually have to contort themselves to bend over backwards to justify that any part of this whole copyright system makes any sense at all.

    Sooner or later, judges will have to clarify what "solely" means with regard to generative work. That definition will most likely suffer from the slippery slope conundrum, where any so-called dividing line can be further divided in half to blur the definition.

    At what point can a fully-automated system be sl

  • From the article:

    In a statement Friday, Thaler's attorney Ryan Abbot said he and his client "disagree with the district court's judgment" and vowed to appeal: "In our view, copyright law is clear that the public is the main beneficiary of the law and this is best achieved by promoting the generation and dissemination of new works, regardless of how they are created."

    Yes, and how does making AI output copyright free not make the public the main beneficiary of the work exactly?
    It is a broken and dishonest view of copyright.
    In copyright the public benefit comes after the benefits for the author. It's main purpose is to allow an author to capitalize on the work before it is absorbed by the public. The public benefit is in the fact that copyright has limits.

To write good code is a worthy challenge, and a source of civilized delight. -- stolen and paraphrased from William Safire

Working...