News organisations led by The New York Times have asked a US court to impose "serious sanctions" on OpenAI, alleging the company lied for two years to conceal its ability to search ChatGPT logs for evidence of copyright infringement. These logs are seen as pivotal to the case, potentially either proving OpenAI's chatbot reproduces paywalled articles verbatim or supporting its defence that the technology is a transformative fair use. The plaintiffs claim OpenAI falsely told the court that searching the logs was infeasible, costly and a threat to user privacy, when it had in fact already conducted such searches before the litigation began.

The alleged deception surfaced during an April re-deposition of OpenAI privacy engineer Vincent Monaco, who reportedly revealed the firm held two de-identified samples of roughly 10 million and 78 million logs that it never disclosed, and had already searched them while researching a filter to block copyrighted content. Plaintiffs say they were instead forced to spend eight months working within a heavily redacted 20-million-log "sandbox" that the court deemed "unusable" after OpenAI used AI to make 19 billion redactions. OpenAI rejects the claims, calling them "blatantly false" and framing the motion as an attempt to invade user privacy as the Times's weakening case forces it to drop claims; the Times counters that its suit has been streamlined and strengthened by adding claims against Microsoft.