NYT: OpenAI Hid Evidence in ChatGPT Copyright Lawsuit

The New York Times reported on June 17, 2024, that OpenAI allegedly concealed evidence in the ongoing copyright lawsuit filed by news publishers. The publishers accuse OpenAI of hiding tools and datasets that could have revealed instances where ChatGPT generated content derived from copyrighted journalistic works. This alleged concealment has led the publishers to file a new motion for sanctions against OpenAI, escalating the legal battle.

The core of the publishers' claim is that OpenAI deliberately obscured its methods for identifying and handling copyrighted material. They contend that the company possesses internal tools and specific datasets that would demonstrate whether ChatGPT outputs were trained on or reproduced copyrighted articles without proper attribution or licensing. The motion for sanctions seeks to penalize OpenAI for what the publishers describe as a deliberate effort to obstruct the discovery process and mislead the court.

This development intensifies the legal dispute, which centers on allegations that OpenAI's AI models, including ChatGPT, were trained on vast amounts of copyrighted text from news organizations without permission. The publishers argue that this unauthorized use constitutes copyright infringement and has harmed their businesses. OpenAI has previously stated its commitment to respecting copyright and has been working to implement measures to address these concerns, but the latest allegations suggest a significant gap between their public statements and their actions during the legal proceedings.

The publishers' legal team is seeking specific remedies as part of the sanctions, though the exact nature of these remedies has not been fully detailed in public filings. The outcome of this sanctions motion could have significant implications for the broader copyright lawsuit and for how AI companies handle training data derived from copyrighted sources. The case is being closely watched by the media industry and the artificial intelligence sector alike, as it addresses fundamental questions about intellectual property in the age of generative AI.