The New York Times and Daily News have escalated their copyright battle against OpenAI, accusing the company of hiding internal tools capable of searching its training data and ChatGPT logs. The publishers are asking a Manhattan federal judge to sanction OpenAI for alleged discovery misconduct.
The dispute sharpened following a deposition of an OpenAI employee, which publishers claim contradicted two years of representations about the company's search capabilities. The news organizations allege OpenAI chose obstruction over releasing evidence relevant to determining whether its systems reproduced copyrighted journalism.
OpenAI has rejected the allegations. Spokesperson Drew Pusateri called them "blatantly false," stating the Times was attempting to invade user privacy. The company maintains it will defend fair use principles and user confidentiality.
The underlying lawsuit, filed in late 2023, alleges OpenAI and Microsoft used copied journalism to develop competing AI products. OpenAI argues AI training on online material qualifies as fair use. Publishers contend their costly original reporting was used without permission or compensation.
The court must now decide if sanctions are warranted. While the ruling won't decide the broader copyright question, it could set precedents for how AI companies handle internal datasets and user records in future litigation.