This is going to go the way of the Silverman case. On quote from that judge:
“This is nonsensical,” he wrote in the order. “There is no way to understand the LLaMA models themselves as a recasting or adaptation of any of the plaintiffs’ books.”
The Silverman case isn’t over. The judge took the position that the output themselves are not infringement, as I think most people agree since it is a transformation, but the core of the case is still ongoing - that the dataset used to train these models contained their copyrighted work. Copying is one of the rights granted to copyright holders and, unlike the Google case a few years back, this is for a commercial product and the books were not legally obtained. Very different cases. I would be surprised if Silverman and the others lost this lawsuit.
Copyright is more about distribution and deprivation than copying.
There is absolutely nothing preventing me from sitting down and handwriting the entirety of the LOTR in calligraphic script.
I can even give that copy to other people, as it is a “derivative work,” and I’m not attempting to profit from it.
There’s not even anything preventing me from scanning every page and creating a .pdf file for personal use, as long as I don’t distribute it.
Hell, the DMCA even allows me to rip a movie as long as I’m keeping it for personal use.
I don’t see anything here that can not be argued against with fair use. The case is predicated upon the idea that if you give it the correct prompts, it’ll spit out large amounts of copyrighted text.
If you were describing that as an interaction with a person, you’d call that coercion and maybe even entrapment.
The intent of the scraping was not explicitly distribution.
This is going to go the way of the Silverman case. On quote from that judge:
The Silverman case isn’t over. The judge took the position that the output themselves are not infringement, as I think most people agree since it is a transformation, but the core of the case is still ongoing - that the dataset used to train these models contained their copyrighted work. Copying is one of the rights granted to copyright holders and, unlike the Google case a few years back, this is for a commercial product and the books were not legally obtained. Very different cases. I would be surprised if Silverman and the others lost this lawsuit.
Just to clarify most people do not agree. A lot of people are explicitly arguing that.
It is with respect to that argument. The claim in question was thrown out.
The remaining claim is unrelated.
Copyright is more about distribution and deprivation than copying.
There is absolutely nothing preventing me from sitting down and handwriting the entirety of the LOTR in calligraphic script.
I can even give that copy to other people, as it is a “derivative work,” and I’m not attempting to profit from it.
There’s not even anything preventing me from scanning every page and creating a .pdf file for personal use, as long as I don’t distribute it.
Hell, the DMCA even allows me to rip a movie as long as I’m keeping it for personal use.
I don’t see anything here that can not be argued against with fair use. The case is predicated upon the idea that if you give it the correct prompts, it’ll spit out large amounts of copyrighted text.
If you were describing that as an interaction with a person, you’d call that coercion and maybe even entrapment.
The intent of the scraping was not explicitly distribution.