OpenAI challenges court order to preserve ChatGPT user data, cites privacy concerns

OpenAI is appealing a US federal court order requiring it to indefinitely retain user-generated ChatGPT outputs, arguing the directive conflicts with its privacy commitments. 

The order stems from a lawsuit filed by The New York Times (NYT), which accuses OpenAI and its key partner Microsoft of copyright infringement involving millions of NYT articles.

The company filed a motion on June 3 asking US District Judge Sidney Stein to vacate the preservation order, which mandates storing and segregating all ChatGPT output logs.

OpenAI CEO Sam Altman criticised the requirement on X , stating, "We will fight any demand that compromises our users' privacy; this is a core principle." He added the request "sets a bad precedent."

The Times’ request is part of a broader legal push to establish that OpenAI and Microsoft unlawfully used copyrighted news content to train artificial intelligence systems, including ChatGPT. 

In a 2023 lawsuit, the newspaper alleged that the tech companies reproduced and regurgitated its reporting without authorization or compensation, citing examples where ChatGPT generated near-verbatim excerpts from paywalled or copyrighted articles.

In April, Judge Stein allowed the case to proceed, rejecting parts of a dismissal motion from OpenAI and Microsoft. He said the Times had made a "plausible" case that the defendants induced users to generate infringing content and that the examples of replicated output were "numerous" and "widely publicized."

OpenAI argues that the preservation order forces it to abandon industry-standard data deletion practices, undermining user trust. 

"The plaintiffs are demanding we retain consumer ChatGPT and API customer data indefinitely—this fundamentally conflicts with the privacy commitments we have made to our users," the company said in a public statement. 

OpenAI noted that any relevant data is already held in a secure legal-hold system and is inaccessible for other purposes.

At present, The New York Times does not have a content licensing or sharing agreement with OpenAI. 

However, OpenAI has signed licensing deals with other major publishers, including The Associated Press, Financial Times, and Germany’s Axel Springer, as it seeks to resolve legal uncertainty around training data through formal partnerships.