On November 20, 2024, The New York Times filed a letter in the Southern District of New York alleging that OpenAI engineers had deleted potential evidence in an ongoing copyright lawsuit.
The lawsuit, initiated in December 2023, accuses OpenAI and Microsoft of using copyrighted materials to train AI models without permission. During the discovery phase, OpenAI provided two virtual machines for the plaintiffs to search for evidence of infringement. The New York Times legal team spent over 150 hours analyzing data on one machine.
However, on November 14, search data from one of the virtual machines was erased by OpenAI engineers. While some data was recovered, critical file names and folder structures were missing. This resulted in the New York Times needing to recreate their research.
The New York Times claims this lost data is crucial evidence, and the deletion represents a serious setback to their case. OpenAI, in turn, attributes the deletion to a “glitch”. The case is still in the discovery phase.
The incident has raised questions about data management protocols during legal proceedings. This event may impact how data is handled in future lawsuits involving large tech companies and AI training data.
