Tensions over Lost Data: OpenAI and Publishers in Legal Dispute

In a sensitive legal dispute between OpenAI and the publishers The New York Times and Daily News, there was an unexpected turn: Allegedly accidentally deleted data by OpenAI could complicate the case. The publishers accuse OpenAI of unauthorized access to their content to train the company's AI models. As part of the proceedings, OpenAI had agreed to provide two virtual machines on which the publishers' lawyers could search for their copyrighted material. However, on November 14, one of the machines was deleted, leading to a loss of valuable search data, as stated in a letter to the court in New York. OpenAI did attempt to restore the data and was mostly successful, but the folder structure and file names were irrevocably lost. This information would have been necessary to trace where the articles were used in the AI models, making it necessary for the publishers to redo their work. Interestingly, the publishers do not suspect any intentional actions on the part of OpenAI. Nevertheless, they emphasize that OpenAI is in the best position to search its data holdings for potential copyright infringements. A statement from OpenAI was absent, but their lawyers responded to the allegations on Friday. They denied the blame and instead attributed the mishap to configuration changes made at the plaintiffs' request. The storage device in question was intended only as a temporary buffer, so no files were actually lost.