A rule against copyrighting AI art will be unworkable.
I’ve generally been against giving AI works copyright, but this article presented what I felt were compelling arguments for why I might be wrong. What do you think?
I haven’t seen anyone that has been able to reproduce complete works from an LLM. Open AI also actively stops people from even trying to reproduce anything that resembles copyrighted materials. Signaling their commercial purpose isn’t to substitute for the plaintiff’s works. Filing suit doesn’t make their claims true, you should hold off on hasty judgements.
Again, it’s not a question of reproducing books in an LLM. The allegation is that the openAI developers downloaded books illegally to train their AI.
You need to pay for your copy of a book. That’s true if you are a student teaching yourself to write, and it’s also true if you are an AI developer training an AI to write. In the latter case, you might also need to pay for a special license.
Is it possible that the openAI developers can bring the receipts showing they paid for each and every book and/or license they needed to train their AI? Sure, it’s possible. If so, the lawyers who brought the suit would look pretty silly for not even bother to check.
But openAI used a whole lot of books, which cost a whole lot of money. So I wouldn’t hold my breath.
When an act potentially implicates copyright but is a necessary step in enabling noninfringing uses, it frequently qualifies as a fair use itself.
Yeah, I think they’ve got a chance. You also definitely don’t need to pay to use books. You can just receive it for free from someone. That’s why college course books make all those revisions and bundle in software to stop people from sharing.
If you are college student, learning to write professionally, is it fair use to download copyrighted books from Z-Library in order to become a better writer? If you are a musician, is it fair use to download mp3s from The Pirate Bay in order to learn about musical styles? How about film students, can they torrent Disney movies as part of their education?
I’m certain that every court in the US would rule that this is not fair use. It’s not fair use even if pirated content ultimately teaches a student how to create original, groundbreaking works of writing, music, and film.
Simply being a student does not give someone free pass to pirate content. The same is true of training an AI, and there are already reports that pirated material is in the openAI training set.
If openAI could claim fair use, then almost by definition The Pirate Bay could claim fair use too.
If the students are using the works for purposes such as analyzing, critiquing, or illustrating a point, and not merely reproducing them, they have a strong case for fair use. That’s all these models are, original analysis of their training data in comparison with each other. This use is more likely to be considered transformative, meaning that they add something new or different to the original work, rather than merely copying it. If you need it said another way, here’s a link to a video about this sort of thing.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@beehaw.org
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
I haven’t seen anyone that has been able to reproduce complete works from an LLM. Open AI also actively stops people from even trying to reproduce anything that resembles copyrighted materials. Signaling their commercial purpose isn’t to substitute for the plaintiff’s works. Filing suit doesn’t make their claims true, you should hold off on hasty judgements.
Again, it’s not a question of reproducing books in an LLM. The allegation is that the openAI developers downloaded books illegally to train their AI.
You need to pay for your copy of a book. That’s true if you are a student teaching yourself to write, and it’s also true if you are an AI developer training an AI to write. In the latter case, you might also need to pay for a special license.
Is it possible that the openAI developers can bring the receipts showing they paid for each and every book and/or license they needed to train their AI? Sure, it’s possible. If so, the lawyers who brought the suit would look pretty silly for not even bother to check.
But openAI used a whole lot of books, which cost a whole lot of money. So I wouldn’t hold my breath.
To quote again:
Yeah, I think they’ve got a chance. You also definitely don’t need to pay to use books. You can just receive it for free from someone. That’s why college course books make all those revisions and bundle in software to stop people from sharing.
Simple question:
If you are college student, learning to write professionally, is it fair use to download copyrighted books from Z-Library in order to become a better writer? If you are a musician, is it fair use to download mp3s from The Pirate Bay in order to learn about musical styles? How about film students, can they torrent Disney movies as part of their education?
I’m certain that every court in the US would rule that this is not fair use. It’s not fair use even if pirated content ultimately teaches a student how to create original, groundbreaking works of writing, music, and film.
Simply being a student does not give someone free pass to pirate content. The same is true of training an AI, and there are already reports that pirated material is in the openAI training set.
If openAI could claim fair use, then almost by definition The Pirate Bay could claim fair use too.
If the students are using the works for purposes such as analyzing, critiquing, or illustrating a point, and not merely reproducing them, they have a strong case for fair use. That’s all these models are, original analysis of their training data in comparison with each other. This use is more likely to be considered transformative, meaning that they add something new or different to the original work, rather than merely copying it. If you need it said another way, here’s a link to a video about this sort of thing.
So you believe that if you download an mp3 and claim you are “analyzing” it, then you can’t be liable for IP infringement?
Wow, I wonder why the Napster defendants never thought of that. They could have saved tens of thousands of dollars.
They were helping people to reproduce and distribute copyrighted works. There’s a world of difference here.