OpenAI faked inability to search training data, hid billions of logs, NYT says

OpenAI faked inability to search training data, hid billions of logs, NYT says. In a Thursday sanctions motion filed by news organizations led by The New York Times, OpenAI is accused of concealing evidence in its dispute over whether ChatGPT was built and used using copyrighted material. The motion points to testimony from OpenAI privacy engineer Vincent Monaco after the court compelled a re-deposition in an April proceeding. NYT alleges Monaco revealed OpenAI misled the court for two years about the cost and burden of searching ChatGPT logs. The filing further claims OpenAI initially portrayed itself as unable to search large anonymized log samples, despite conducting such searches before litigation. OpenAI denies bad faith, saying the request is aimed at accessing more logs and alleges privacy violations by competitors.