New York Times-led newspapers ask court to sanction OpenAI, alleging it hid its ability to search for copied articles and deleted ChatGPT conversations

What happened

On 9 July 2026, a group of newspapers including The New York Times and the New York Daily News asked the federal court in Manhattan to sanction OpenAI in the consolidated copyright litigation over the use of their journalism to train ChatGPT, Reuters reported. In their filing, the newspapers said OpenAI falsely told the court it could not search its large language models and output logs for their copyrighted material, while hiding that it had performed such searches "even before the first News Plaintiff filed suit." They also alleged OpenAI deleted billions of relevant ChatGPT conversations or made them unsearchable. The sanctions they seek include attorneys' fees and a court finding that OpenAI's chat logs showed the company misused their copyrighted works. The filing cites later testimony by an OpenAI employee that the company had "performed multiple searches for News Plaintiffs' content." OpenAI spokesperson Drew Pusateri rejected the allegations as "blatantly false," saying the newspapers were persisting in efforts to invade users' privacy as their case weakens; OpenAI has previously argued that turning over ChatGPT conversation logs would risk violating users' privacy. The Times' lead attorney Ian Crosby said OpenAI had claimed for over two years that searching ChatGPT outputs was "infeasible, burdensome, and invasive of users' privacy" while concealing that it had already done such searches.

Why it matters to creators

The motion is about the evidence that could prove or disprove how publishers' journalism was used in training and outputs. If the court accepts the newspapers' account, the dispute shifts from whether AI training is fair use to whether the AI company concealed and destroyed the records creators need to show their work was copied. The requested remedy — a court finding that the chat logs showed misuse of the newspapers' works — would, if granted, establish a key liability fact for the plaintiff publishers without a trial on that point. It also puts OpenAI's privacy-based resistance to producing chat logs at the centre of the case, a tension affecting any creator whose work may appear in model outputs.

What is not yet known

The sanctions allegations are contested and no court ruling on the motion is reported in these sources; OpenAI denies them. It is not yet known whether the court will impose any sanction, award fees, or make the requested finding about the chat logs, or how the motion interacts with the parties' summary-judgment motions filed in September 2026.