New York Times and Daily News accuse OpenAI of withholding evidence in ChatGPT copyright case

The New York Times and The Daily News have filed a motion for sanctions against OpenAI, alleging the company concealed internal tools and datasets that could reveal whether ChatGPT outputs reproduced their copyrighted journalism. The lawsuit, now in its second year, accuses OpenAI of training its generative AI models on the outlets’ content without permission.

According to court filings, OpenAI previously argued it lacked the technical ability to search its training corpus or ChatGPT chat logs, citing burdensomeness and user-privacy concerns. However, plaintiffs cite the testimony of OpenAI data privacy engineer Vinnie Monaco, who allegedly revealed in an April deposition that the company had already conducted internal searches of its training corpus to identify copyrighted journalism works.

Plaintiffs further allege OpenAI maintained a database of approximately 78 million de-identified ChatGPT conversations, which it used internally to assess infringement. They also claim OpenAI implemented a tool called “Project Giraffe,” which included a “Bloom” filter to detect and record instances of output regurgitation after the lawsuit was filed.

The outlets allege OpenAI deleted billions of ChatGPT outputs after the lawsuit began, in violation of a court preservation order, and substituted millions of logs in a requested sample of 20 million chat logs. The sample, which plaintiffs negotiated down from 120 million, was ultimately deemed “unusable” by the court due to excessive redactions.

The plaintiffs are asking the judge to prevent OpenAI from using the disputed 20 million chat log sample as evidence, accept as fact that ChatGPT logs would have shown major regurgitation of their content, bar OpenAI from arguing that the logs do not demonstrate substantial regurgitation, and award legal fees incurred while pursuing the evidence.