“Taken as true, these facts give rise to a plausible inference that defendants at a minimum had reason to investigate and uncover end-user infringement,” Stein wrote.
OpenAI’s ability to maintain an “ongoing relationship” OpenAI’s claim that ChatGPT’s outputs are not contributory to infringement is also supported by the fact that ChatGPT provides outputs that respond directly to user prompts. “substantial noninfringing uses” Exonerative.
OpenAI defeats some claims
Stein’s decision is likely to disappoint OpenAI, even though Stein did drop some NYT claims.
Likely upsetting for news publishers, including a “free-riding” ChatGPT is a time-sensitive service that profits unfairly from its use. “hot news” Items, including the NYT Wirecutter postings. Stein explained that news publishers had failed to plausibly claim non-attribution, which is crucial to a free rider claim, because ChatGPT, for example cites the New York Times when sharing information from Wirecutter postings. Stein wrote that the Copyright Act preempts these claims, and thus, OpenAI’s motion for dismissal.
Stein also dismissed an allegation from the NYT about alleged removals of copyright management (CMI), Stein said that this cannot be proven because ChatGPT reproduces NYT articles with CMI removed.
Stein said that the Digital Millennium Copyright Act requires news publishers to prove that ChatGPT outputs are “closely identical” to original works, and allows publishers’ claims based upon excerpts. “would risk boundless DMCA liability”Use of block quotes or any other form of quotation without CMI is prohibited.
OpenAI’s spokesperson refused to comment on the ruling. Instead, he reiterated OpenAI’s long-held position that AI training on copies of copyrighted works constitutes fair use. OpenAI warned Donald Trump last month that the US would lose out in the AI race if courts ruled that this argument was invalid.
“ChatGPT helps enhance human creativity, advance scientific discovery and medical research, and enable hundreds of millions of people to improve their daily lives,” OpenAI’s spokesperson stated. “Our models empower innovation, and are trained on publicly available data and grounded in fair use.”
