The New York Times and The Daily News have escalated their two-year-old copyright lawsuit against OpenAI, accusing the AI company of concealing evidence about its ability to search customer chat logs and training datasets for their copyrighted works. In a court filing, the news outlets allege that OpenAI lied when it claimed it lacked the technical means to search its own training corpus, and that it had secretly amassed a database of roughly 78 million de-identified ChatGPT conversations to track infringement.
The latest salvo centers on a deposition given in April by Vinnie Monaco, OpenAI’s data privacy engineer. According to the plaintiffs, Monaco revealed that OpenAI had already conducted internal searches and evaluations of its training corpus for copyrighted journalism works — directly contradicting the company’s repeated assertions that such searches were technically burdensome or impossible. The deposition also allegedly disclosed that, before the Times filed its lawsuit in late 2023, OpenAI had built a searchable database of about 78 million de-identified ChatGPT conversations, which it used internally to assess how often it was reproducing others’ works.
Also read: Paris-based voice AI startup Gradium raises $100M seed round, backed by Nvidia
Project Giraffe and the Bloom Filter
The plaintiffs further claim that shortly after the lawsuit was filed, OpenAI implemented a tool called the “Bloom” filter as part of a broader internal initiative known as “Project Giraffe.” This filter was designed to detect and log instances of “regurgitation” — when ChatGPT reproduces copyrighted text verbatim or near-verbatim in its outputs. The existence of this system, the plaintiffs argue, proves that OpenAI was actively monitoring for infringement while simultaneously arguing in court that it could not do so.
These revelations are particularly significant because they undercut OpenAI’s central defense in the discovery process. The plaintiffs had originally requested a sample of 120 million chat logs to determine how often ChatGPT reproduced their journalism. OpenAI negotiated that number down to 20 million logs, which it finally submitted in December. However, the plaintiffs claim the sample was so heavily redacted that a judge called it “unusable.” They also allege that OpenAI deleted billions of ChatGPT outputs after the lawsuit was filed, in direct violation of a court-issued preservation order, and that it substituted millions of logs in the requested sample.
Also read: Meta pulls controversial Instagram AI feature after backlash over privacy and misuse concerns
“If OpenAI genuinely believed that copying our clients’ journalism was fair and legal, it wouldn’t have hid the truth about having done it,” Ian B. Crosby, lead counsel for the plaintiffs, said in a statement.
Sanctions Sought
The news outlets are now asking the court to impose several sanctions on OpenAI. They want the judge to prevent OpenAI from using the 20 million chat log sample as evidence, arguing it is unreliable. They are also asking the court to accept as fact that ChatGPT logs would have shown major regurgitation of the plaintiffs’ content, and to bar OpenAI from arguing that its provided logs do not demonstrate substantial copying. Finally, they are seeking an order requiring OpenAI to pay legal fees incurred while chasing down the allegedly hidden evidence.
OpenAI has pushed back forcefully. In a statement, spokesperson Drew Pusateri denied the allegations, framing them as a desperate move by the Times as its case weakens. “As the Times’ case weakens and they’ve been forced to drop claims against us, they’re persisting with their efforts to invade the privacy of people who have nothing to do with this case, including by making these blatantly false allegations,” Pusateri said. “We’ll continue defending our users’ privacy and the long-established principles of fair use.”
The dispute is the latest flashpoint in a broader legal battle that has become a bellwether for the AI industry. A ruling against OpenAI on these discovery issues could set a precedent for how courts handle claims of copyright infringement in the training of large language models, where the scale of data makes traditional discovery methods uniquely challenging.
Frequently Asked Questions
What is the core allegation in the New York Times lawsuit against OpenAI?
The New York Times and The Daily News allege that OpenAI violated copyright law by training its AI models on their articles without permission and reproducing that journalism in ChatGPT outputs.
What evidence do the plaintiffs claim OpenAI hid?
They claim OpenAI hid internal searches of its training corpus for copyrighted works, and a database of about 78 million de-identified ChatGPT conversations used to track infringement, which contradicts OpenAI’s earlier claims that such searches were technically burdensome.
What is Project Giraffe?
Project Giraffe was a set of internal tools at OpenAI that included a ‘Bloom’ filter designed to detect and record instances of ChatGPT reproducing copyrighted content, allegedly implemented shortly after the lawsuit was filed.
What sanctions are the plaintiffs seeking?
The plaintiffs are asking the court to prevent OpenAI from using its 20 million chat log sample as evidence, to accept as fact that ChatGPT logs would have shown major regurgitation of their content, and to require OpenAI to pay legal fees.
How has OpenAI responded to these new allegations?
OpenAI spokesperson Drew Pusateri denied the allegations, calling them ‘blatantly false’ and accusing the Times of trying to invade user privacy as its own case weakens.
Be the first to comment