NYT accuses OpenAI of hiding evidence in copyright lawsuit
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI data privacy engineer Vinnie Monaco allegedly revealed in an April deposition that the company had already run internal searches of its training corpus to locate copyrighted journalism before the NYT filed suit
- OpenAI had compiled a database of roughly 78 million de-identified ChatGPT conversations it was using internally to measure infringement, per Monaco's testimony, despite arguing in court that producing such data would be technically burdensome
- OpenAI built a "Bloom" filter inside a toolset called "Project Giraffe" to detect and log regurgitation of content shortly after the lawsuit was filed, according to the plaintiffs
- The New York Times and Daily News claim OpenAI deleted billions of ChatGPT outputs after the suit was filed in violation of a court preservation order and substituted millions of logs in the sample it did produce
- OpenAI negotiated the plaintiffs' request for 120 million chat logs down to 20 million, then submitted a version so heavily redacted that the court called it "unusable"
- The plaintiffs are asking the judge to bar OpenAI from using the 20 million log sample as evidence, accept as fact that logs would show major regurgitation, block OpenAI from arguing the logs don't show infringement, and award legal fees
- OpenAI spokesperson Drew Pusateri denied the allegations, accusing the Times of "blatantly false" claims and saying it is "invading the privacy of people who have nothing to do with this case"
Why it matters: If the court grants the plaintiffs' requests, OpenAI could be barred from defending itself with its own chat log sample and have key infringement findings accepted as established fact — effectively conceding major elements of the case without a merits trial, while owing the publishers' legal fees on top of any damages.
