Lawsuits are never a lovefest, but the copyright battle between The Fresh York Times and OpenAI and Microsoft is becoming particularly contentious. This week, the Los Angeles Times alleged that OpenAI engineers inadvertently removed data that the writing team spent more than 150 hours extracting as potential evidence.
OpenAI managed to recover most of the data, but the Times’ legal team says the original file names and folder structure are still missing. According to A declaration filed Wednesday in court by Jennifer B. Maisel, a lawyer for the newspaper, means the information “cannot be used to determine where copied articles by news writers” may have been incorporated into OpenAI’s artificial intelligence models.
“We disagree with the descriptions provided and will provide our response shortly,” OpenAI spokesman Jason Deutrom told WIRED in a statement. The Fresh York Times declined to comment.
Last year, The Times filed a copyright lawsuit against OpenAI and Microsoft, alleging that the companies illegally used its articles to train artificial intelligence tools such as ChatGPT. The case is one of many ongoing legal battles between artificial intelligence companies and publishers, including a similar lawsuit filed by the Daily News led by some of the same lawyers.
The Times’ case is currently ongoing, meaning both sides are turning over requested documents and information that could become evidence. As part of that lawsuit, the court required OpenAI to show the Times its training data, which is a gigantic deal – OpenAI has never publicly disclosed exactly what information was used to build its AI models. To reveal this, OpenAI created what the court calls a “sandbox” consisting of two “virtual machines” that Los Angeles Times lawyers could search. In her statement, Maisel said OpenAI engineers “deleted” data organized by the Times team on one of those machines.
According to Maisel’s filing, OpenAI admitted that the information had been deleted and attempted to resolve the issue shortly after receiving the notification earlier this month. But when the newspaper’s lawyers looked at the “restored” data, they found it was too disorganized, forcing them to “recreate their work from scratch, using significant man-hours and computer processing time,” several other Times lawyers said in letter submitted to the judge on the same day as Maisel’s statement.
Lawyers noted that they had “no reason to believe” the removal was “intentional.” In emails submitted as evidence along with Maisel’s letter, OpenAI adviser Tom Gorman referred to data deletion as a “glitch”.