United States of America: Southern District of New York District Court ordered OpenAI produce 20 million ChatGPT logs (The New York Times v OpenAI, Microsoft)

On 2 December 2025, the United States District Court for the Southern District of New York ordered OpenAI to produce 20 million ChatGPT logs within 7 days after denying the company's motion for reconsideration. The case involves a dispute between OpenAI and the New York Times, who, alongside other media organisations, claim that OpenAI used copyrighted works for AI training purposes. The order aims to uncover information necessary to evaluate whether copyright infringement occurred and follows a lengthy conflict as the parties negotiated how much user data would be necessary for OpenAI to produce. The Court noted, first, that the logs were relevant to the case not only to the extent that they contained reproductions of New York Times works, but also to OpenAI's fair use and non-infringing uses defences. Second, it noted that the sample is proportional, as it represents less than 0.05% of the total logs retained by OpenAI and the burden of production is minimal. Third, the Court recognised that privacy concerns of OpenAI users are sincere, but notes that privacy is only one factor in the proportionality analysis. Finally, the Court concludes that OpenAI has failed to demonstrate how users' privacy rights are not adequately protected by the combination of OpenAI's internal de-identification tool, the reduction of the data request to 20 million logs, and a protective order on the logs.