The court ordered Microsoft and OpenAI to disclose internal documents about data collection

The court ordered Microsoft and OpenAI to publish internal correspondence regarding methods of data collection for training neural networks. Previously, the companies had tried to hide these documents, citing confidentiality, but the court decision led to their disclosure as part of an antitrust lawsuit.

In the published letters, company executives privately called the mass parsing of news content an "amazing theft" and a "death loop" for the publishing industry. These statements directly contradict the corporations' public statements that the use of copyrighted materials falls under the doctrine of fair use.

Internal discussions highlight the concerns of top management regarding the long-term sustainability of current training models. One high-ranking employee characterized the data collection process as the "largest theft of labor in human history", which raises questions about the legal purity of the training samples on which modern LLMs are based.

The disclosure of correspondence has become a turning point in a series of lawsuits from media holdings and copyright holders. The documents that have become available to the public can now be used as evidence in ongoing proceedings on copyright infringement and unfair competition.

This incident sets a precedent for future proceedings in the field of AI. The publication of correspondence discussed on Hacker News forces tech giants to reconsider their defense strategy in court and may lead to stricter requirements for transparency of data sources for training neural networks.