OpenAI responds publicly to New York Times demand for user chat data

OpenAI published a public statement objecting to the scope of a magistrate judge’s order requiring it to preserve all ChatGPT output logs as evidence in The New York Times’ copyright suit, rather than confining its objection to court filings. Chief operating officer Brad Lightcap wrote that the company “strongly believe[d] this is an overreach” by the Times, arguing that indefinite retention of deleted and temporary chats put user privacy at risk without materially helping resolve the underlying copyright dispute.

Lightcap said the order affected the large majority of ChatGPT’s consumer and API users — Enterprise and education accounts, which are typically covered by stronger contractual data-handling terms, were excluded — and that OpenAI had already filed to have the order narrowed or reversed. He framed the fight as one over user trust as much as legal strategy, describing OpenAI’s ongoing work to build privacy protections into its products and arguing the preservation requirement conflicted with those commitments and with data-protection norms in the US and Europe.

The Times had sought the logs on the theory that some users’ conversations would show them prompting ChatGPT to reproduce paywalled Times content, which it argued was relevant to proving the underlying infringement claims against OpenAI. OpenAI’s public framing — that a court order in a corporate copyright case was compelling disclosure of ordinary users’ private conversations — was a deliberate attempt to shift the terms of the dispute in the court of public opinion as well as in court, and it set the pattern for a running argument between the two companies over data scope that continued through the rest of the year, including a subsequent OpenAI complaint that the Times’ demand had grown to millions of conversations.