Judge Gonzalez-Rogers, who presided over the Musk v. Altman trial, has another big case: Hendrix v. Apple.
Notably, in her order this week, Judge Gonzalez-Rogers “further instructed to discuss and consider whether earlier summary judgment motions on the core issue of fair use are possible as part of the scheduling decision.”
Apple already teed up the issue of fair use in its Answer. (1) “Apple admits that it used a subset of the RedPajama dataset called ‘Books,’ to train OpenELM, which is a non-commercial research model that is not part of Apple Intelligence. Apple admits that the datasets used to train certain AFM models contained books or portions thereof.”
(2) “The conduct at issue in this lawsuit is fair use, not copyright infringement. Large language models (‘LLMs’) are trained on datasets containing many billions of words of text, a practice also used for Apple’s OpenELM research models and Apple Foundation Models (“AFM”). Applying copyright law’s use-specific analysis here, the evidence will show that Apple’s use was fair. If accepted, Plaintiffs’ rigid application of copyright law—imposing liability on the use of copyrighted works to advance research and technological innovation—would ‘stifle the very creativity which that law is designed to foster.’”
Related Stories
