A Manhattan federal court judge, Ona Wang, has ordered OpenAI to disclose millions of anonymized conversation records from its ChatGPT platform. The decision is related to demands made by The New York Times and other media organizations in lawsuits alleging copyright infringement.
According to reports, the plaintiffs accuse OpenAI and Microsoft of developing artificial intelligence models using copyrighted materials without permission, stating that ChatGPT's distorted reproduction of articles undermines the value of professional journalism. Although the company argued that disclosing user data would violate trade secrets and privacy principles, the court did not accept these arguments.
According to the ruling, OpenAI must provide approximately 20 million conversation records. These records will be presented with personally identifiable information removed, but the content of the prompts and responses will be preserved. Experts note that these volumes are expected to play a crucial role in legal investigations to determine the extent to which ChatGPT replicates protected texts.
At the heart of the dispute lies the application of "fair use" principles in artificial intelligence training. While OpenAI presents its use of a vast dataset collected from the internet as a transformative activity, media organizations claim it constitutes unauthorized use and, at times, direct reproduction of source texts. The increasing number of similar lawsuits in the US and Canada indicates growing pressure on the AI industry's data policies.