OpenAI accidentally erases evidence of using newspaper content to train AI models: Report

Though OpenAI admitted to the error as a “glitch” and even tried to recover the data, whatever it managed to salvage can no longer be used as evidence in court

OpenAI accidentally erased important evidence gathered by major newspapers including the New York times in their lawsuit regarding training data, according to a report by The Verge which cited a court filing.

OpenAI CEO Sam Altman speaks during the Microsoft Build conference at the Seattle Convention Center Summit Building in Seattle, Washington. (AFP)

This comes after the newspapers' legal teams spent over 150 hours searching through OpenAI's training data to find cases where their news articles were included, according to the report.

Though OpenAI admitted to the error as a “glitch” and even tried to recover the data, whatever it managed to salvage was incomplete as well as unreliable, meaning it can no longer be used as evidence in court, the report read.

Even the New York Times’ attorneys were quoted as saying that they had “no reason to believe” it was intentional.

What is the lawsuit against OpenAI by the New York Times about?

The New York Times initiated the legal case last December, claiming OpenAI and Microsoft built their AI tools, “copying and using millions” of the its news articles, making it liable for “billions of dollars in statutory and actual damages.”

The publication has already spent over $1 million fighting with OpenAI in court, an expense very few news outlets can bear.

However, this also comes at a time when OpenAI had made deals with other major publishers like Axel Springer, Conde Nast, and Vox Media amid the same concerns regarding content