An unusual setback has occurred in the lawsuit against OpenAI: the company has just removed some of the work of lawyers representing the opposing side.
The lawsuit was filed by a group of news organizations, including The New York Times and The Chicago Tribune. The lawsuit accuses OpenAI of using their articles for training and infringing their copyrights.
Lawyers for the news organization have been examining OpenAI’s training data under a tightly controlled discovery process involving computers without internet connections.
According to legal documents, the error occurred last week.
“Since November 1, 2024, Plaintiff News has spent over 150 man-hours searching OpenAI’s training data for examples of Plaintiff News’s Asserted Work. Plaintiff News stores the results of their searches in two dedicated virtual machines provided by OpenAI,” the documents reviewed by Business Insider said.
“On November 14, 2024, Plaintiff News learned that OpenAI engineers had deleted all of Plaintiff News’ programs and search result data stored on one of its dedicated virtual machines.”
It goes on to say that OpenAI managed to recover some of the data, but failed to recover the structure or names of the files, which the lawyers say makes it useless.
It’s unclear how the data on the server was wiped, but in a second legal filing, an attorney for the newspaper said they had “no reason to believe it was intentional.”
The newspaper’s lawyers asked the judge to have OpenAI redo their search so they don’t have to redo all their (expensive) work.
