There are many upsetting things about the ai industry but digitising purchased material ain’t it.
If anything this is a boon for information archivism. Unless we ww ourself
Into a new stone age this dataset could easily survive some of the books it contains.
It sounds like the court ruled it fair use at least in part due to the fact they destroyed the copies after digitizing them too. If this clickbait upsets you, get mad at IP law, not anthropic.
Seems like bad legislation decisions though. Maybe write in a clause that says if you upload a book to train an AI, then once completed you have to get rid of the book, in a manner than can include donating them to libraries or charities.
Any way it goes it’s a loss. Why waste the paper, glue, ink and such. Would be great if they created a database when they uploaded each book and shared it to the world with direct purchase of the digital copy to the owner of the work. So the other 30 AIs that come along can just download them there, and they already know a set price, so if we see the company doesn’t pay at least that much, we know they are stealing the works
The ‘pirating’ news from a couple of months ago was Meta, specifically. But I’m sure Anthropic did some too.
The issue I’ve always had wasn’t that they didn’t own a copy to read/reference. It’s that they’re effectively creating derivative works from that content, which they haven’t licensed for that use.
According to my understanding of copyright law (IANAL but I took a few IP law classes on in college) every author whose work was fed into that beast could have an argument that they share copyright in the derivative work that comes out of it.
There was actually just a big ruling on a case involving this, here’s an article about it. In short: a judge granted summary judgment that establishes that training an AI does not require a license or any other permission from the copyright holder, that training an AI is not a copyright violation and they don’t hold any rights over the resulting model.
I’m assuming this case is why we have this news about Anthropic scanning books coming out right now too.
Funny, for me it was quite heartening. If it had gone the other way it could have been disastrous for freedom of information and culture and learning in general. This decision prevents big publishers like Disney from claiming shares of the pie - their published works are free for anyone with access to them to train on, they don’t need special permission or to pay special licensing fees.
As a photographer and the spouse of a writer, they are making massive profits off of a product that wouldn’t exist if they didn’t train it. By the very way the technology works, there’s a little bit of our work scattered in everything they do. If I included a sample of a piece of music in a song I recorded, or included a copyrighted painting in the background if a movie I was making, is would have to get a license. Why is this any different?
They should have done something more like a commodity license as it exists in music:
The composer of a song cannot prevent a new artist from recording a cover of their music if it has been previously released. The original composer is legally forced to grant them a license (hence “compulsory license”). But that license is at a pre-negotiated minimal rate. The new artist is free to try to negotiate a lower rate if the composer agrees. But the original composer can’t stop the new artist from recording a cover. And the new artist has to pay them for it.
Unfettered access is granted and the composer gets their share. Win-win.
The judgment in the article I linked goes into detail, but essentially you’re asking for the law to let you control something that has never been yours to control before.
If an AI generates something that does indeed provably contain a sample of a piece of music in a song you recorded, then yes, that output may be something you can challenge as a copyright violation. But if the AI’s output doesn’t contain an identifiable sample, then no, it’s not yours. That’s how copyright works, it’s about the actual tangible expression.
It’s not about the analysis if copyrighted works, which is what AI training is doing. That’s never been something that copyright holders have any say over.
So, people were angry at them for pirating books. Now we find they actually purchased books to scan, and people are angry about that too.
This is rage bait.
There are many upsetting things about the ai industry but digitising purchased material ain’t it.
If anything this is a boon for information archivism. Unless we ww ourself Into a new stone age this dataset could easily survive some of the books it contains.
Yeah, I can’t be mad about this. They bought the books, they can do with the physical media whatever they like.
It sounds like the court ruled it fair use at least in part due to the fact they destroyed the copies after digitizing them too. If this clickbait upsets you, get mad at IP law, not anthropic.
Seems like bad legislation decisions though. Maybe write in a clause that says if you upload a book to train an AI, then once completed you have to get rid of the book, in a manner than can include donating them to libraries or charities.
Any way it goes it’s a loss. Why waste the paper, glue, ink and such. Would be great if they created a database when they uploaded each book and shared it to the world with direct purchase of the digital copy to the owner of the work. So the other 30 AIs that come along can just download them there, and they already know a set price, so if we see the company doesn’t pay at least that much, we know they are stealing the works
The ‘pirating’ news from a couple of months ago was Meta, specifically. But I’m sure Anthropic did some too.
The issue I’ve always had wasn’t that they didn’t own a copy to read/reference. It’s that they’re effectively creating derivative works from that content, which they haven’t licensed for that use.
According to my understanding of copyright law (IANAL but I took a few IP law classes on in college) every author whose work was fed into that beast could have an argument that they share copyright in the derivative work that comes out of it.
There was actually just a big ruling on a case involving this, here’s an article about it. In short: a judge granted summary judgment that establishes that training an AI does not require a license or any other permission from the copyright holder, that training an AI is not a copyright violation and they don’t hold any rights over the resulting model.
I’m assuming this case is why we have this news about Anthropic scanning books coming out right now too.
That’s disappointing to say the least. I’m sure there will be a few more lawsuits as big publishers like Disney try to get their share of the pie.
Funny, for me it was quite heartening. If it had gone the other way it could have been disastrous for freedom of information and culture and learning in general. This decision prevents big publishers like Disney from claiming shares of the pie - their published works are free for anyone with access to them to train on, they don’t need special permission or to pay special licensing fees.
As a photographer and the spouse of a writer, they are making massive profits off of a product that wouldn’t exist if they didn’t train it. By the very way the technology works, there’s a little bit of our work scattered in everything they do. If I included a sample of a piece of music in a song I recorded, or included a copyrighted painting in the background if a movie I was making, is would have to get a license. Why is this any different?
They should have done something more like a commodity license as it exists in music:
The composer of a song cannot prevent a new artist from recording a cover of their music if it has been previously released. The original composer is legally forced to grant them a license (hence “compulsory license”). But that license is at a pre-negotiated minimal rate. The new artist is free to try to negotiate a lower rate if the composer agrees. But the original composer can’t stop the new artist from recording a cover. And the new artist has to pay them for it.
Unfettered access is granted and the composer gets their share. Win-win.
The judgment in the article I linked goes into detail, but essentially you’re asking for the law to let you control something that has never been yours to control before.
If an AI generates something that does indeed provably contain a sample of a piece of music in a song you recorded, then yes, that output may be something you can challenge as a copyright violation. But if the AI’s output doesn’t contain an identifiable sample, then no, it’s not yours. That’s how copyright works, it’s about the actual tangible expression.
It’s not about the analysis if copyrighted works, which is what AI training is doing. That’s never been something that copyright holders have any say over.