NYT Claims OpenAI Hid Evidence of Copyright Infringement

The New York Times and The Daily News are accusing OpenAI of lying about its ability to search its training data and chat logs for their copyrighted content. This is the latest twist in a two-year lawsuit where the outlets claim OpenAI trained its AI models on their articles and reproduced their journalism. OpenAI previously argued it couldn’t easily search its massive dataset, but a court-ordered deposition allegedly revealed the company had already built tools to do exactly that. A data privacy engineer reportedly admitted OpenAI had an internal database of 78 million de-identified chat logs and a filter called “Bloom” to detect regurgitation. The plaintiffs say OpenAI made it impossible to get reliable evidence by heavily redacting a sample of 20 million logs and possibly deleting billions of outputs after a preservation order. Why it matters: This case could set a precedent on how AI companies handle copyright disputes and discovery—essentially testing whether tech giants can hide behind technical loopholes when accused of copying creative work.