Judge calls out OpenAI’s “straw man” argument in New York Times copyright suit
Apr 04, 2025 - arstechnica.com
A court ruling by Stein found that OpenAI could potentially be liable for contributory infringement due to its ongoing relationship with users and the outputs generated by ChatGPT. Despite OpenAI's argument that ChatGPT has substantial noninfringing uses, Stein suggested that OpenAI should have investigated potential end-user infringement. However, Stein dismissed some claims from the New York Times (NYT), including a "free-riding" claim related to ChatGPT's use of time-sensitive news items, as well as a claim regarding the removal of copyright management information. Stein noted that these claims were pre-empted by the Copyright Act and that the Digital Millennium Copyright Act requires outputs to be "close to identical" to the original work to prove infringement.
OpenAI expressed disappointment with the ruling but emphasized its stance that AI training on copyrighted works constitutes fair use. An OpenAI spokesperson reiterated the company's position that ChatGPT enhances creativity, scientific discovery, and daily life improvements, arguing that their models are grounded in fair use and trained on publicly available data. OpenAI also highlighted the potential negative impact on the US's competitive position in AI if courts rule against their fair use argument.
Key takeaways:
Stein ruled that OpenAI's ongoing relationship with users supports contributory infringement claims despite OpenAI's argument of substantial noninfringing uses.
Stein dismissed some of NYT's claims, including a "free-riding" claim and a claim regarding the removal of copyright management information.
Stein noted that news publishers failed to plausibly allege non-attribution, which is key to a free-riding claim, as ChatGPT cites the NYT when sharing information.
An OpenAI spokesperson reiterated that AI training on copyrighted works is fair use, emphasizing the benefits of ChatGPT in enhancing creativity and innovation.
