NYT Seeks Sanctions, Says OpenAI Hid ChatGPT Log Search Evidence for 2 Years

3 articles · Updated · Ars Technica · Jul 9

The New York Times and 16 other news organizations told a federal court that OpenAI repeatedly lied for years about whether it could search anonymized ChatGPT logs for evidence users used the bot to bypass paywalls.
An April redeposition of OpenAI privacy engineer Vincent Monaco allegedly showed the company had already run such large-scale log searches before litigation, undercutting earlier claims that the process was too costly or technically burdensome.
The publishers said that concealment blocked highly relevant infringement evidence, prolonged discovery, raised costs and burdened the court, and they asked for “serious sanctions” in the Southern District of New York.
OpenAI rejected the accusations as false, arguing the motion is a late attempt to access more user logs, invade privacy and salvage a weakening case after the Times dropped some claims.
The dispute cuts to a central question in the copyright suits: whether ChatGPT regurgitated protected news content often enough to support infringement claims or whether OpenAI can still frame its use as fair use.