Artificial intelligence firm Anthropic hits out at copyright lawsuit filed by music publishing corporations, claiming the content ingested into its models falls under ‘fair use’ and that any licensing regime created to manage its use of copyrighted material in training data would be too complex and costly to work in practice
GenAI tools ‘could not exist’ if firms are made to pay copyright::undefined
The LCA principles also make the careful and critical distinction between input to train an LLM, and output—which could potentially be infringing if it is substantially similar to an original expressive work.
from your second link. I don’t often see this brought up in discussions. The problem of models trained on copyrighted info is definitely different than what you do with that model/output from it. If you’re making money from infringing, the fair use arguments are historically less successful. I have less of an issue with the general training of a model vs. commercial infringing use.
It doesn’t really, I was just kind of restating what you quoted. Since no one factor of fair use is more important than the others, and it is possible to have a fair use defense even if you do not meet all the criteria of fair use, do you have data to back up your claims about moneymaking infringement?
Cool. Thanks. I can see it now.
No, not really, just the pieces over time I’ve read on what wins fair use protections when challenged often talk about the interpretations involved and that profit making was generally seen as detracting from gaining fair use protections when the extent of the transformative nature was in question.
Noncommercial use is more likely to be deemed fair use than commercial use, and the statute expressly contrasts nonprofit educational purposes with commercial ones. However, uses made at or by a nonprofit educational institution may be deemed commercial if they are made in connection with content that is sold, ad-supported, or profit-making. When the use of a work is commercial, the user must show a greater degree of transformation (see below) in order to establish that it is fair.
from your second link. I don’t often see this brought up in discussions. The problem of models trained on copyrighted info is definitely different than what you do with that model/output from it. If you’re making money from infringing, the fair use arguments are historically less successful. I have less of an issue with the general training of a model vs. commercial infringing use.
You’re responsible for infringing works, whether you used Photoshop, copy & paste, or a generative model.
I don’t disagree with that statement. I’m having trouble seeing how that fits with what I said, though. Can you elaborate?
It doesn’t really, I was just kind of restating what you quoted. Since no one factor of fair use is more important than the others, and it is possible to have a fair use defense even if you do not meet all the criteria of fair use, do you have data to back up your claims about moneymaking infringement?
Cool. Thanks. I can see it now. No, not really, just the pieces over time I’ve read on what wins fair use protections when challenged often talk about the interpretations involved and that profit making was generally seen as detracting from gaining fair use protections when the extent of the transformative nature was in question.
This mentions it, but of course it isn’t data on what has been granted protections vs. denials of protection. Harvard counsel primer on copyright and fair use