New York Times Accuses OpenAI of Hiding Evidence in ChatGPT Copyright Trial

The New York Times and a group of newspaper publishers have accused OpenAI of withholding evidence in the ongoing U.S. copyright lawsuit over ChatGPT’s training data, asking a federal court to impose sanctions on the AI company. The publishers allege OpenAI misrepresented its ability to search training datasets and ChatGPT logs for copyrighted news content, while OpenAI has denied the allegations and says it is protecting user privacy and complying with its legal obligations.

The dispute marks the latest escalation in one of the most closely watched legal battles over how generative AI models are trained and whether the use of copyrighted material constitutes fair use under U.S. copyright law.

Publishers Seek Court Sanctions

In a new court filing, the newspaper group—including The New York Times and the New York Daily News—asked the court to sanction OpenAI for what it describes as discovery misconduct.

The publishers allege that OpenAI:

• Misled the court about its ability to search training datasets.
• Previously conducted internal searches for copyrighted journalism.
• Withheld evidence relevant to the lawsuit.
• Failed to preserve or produce certain ChatGPT-related records.

The plaintiffs argue these actions hindered their efforts to determine whether their copyrighted articles were used to train ChatGPT or reproduced in its responses.

OpenAI Rejects the Allegations

OpenAI has denied the accusations, calling them false and misleading.

According to the company:

• It has sought to protect user privacy.
• Broad disclosure of chat logs could expose sensitive customer information.
• It has complied with its legal obligations throughout the litigation.

OpenAI continues to maintain that its AI training practices are protected under the fair use doctrine.

Background of the Copyright Case

The lawsuit, originally filed in 2023, accuses OpenAI and its partner Microsoft of using millions of copyrighted news articles to train generative AI models without authorization.

The publishers argue that AI-generated summaries and responses can compete with original journalism by reducing traffic to news websites and diminishing licensing opportunities.

OpenAI disputes these claims and argues that its models learn general language patterns rather than serving as substitutes for the original works.

Why the Case Matters

The outcome could shape future rules governing AI development.

Key issues before the court include:

• Fair use of copyrighted material.
• AI training data practices.
• Evidence preservation requirements.
• User privacy protections.
• Publisher licensing rights.
• Future regulation of generative AI.

The case is widely viewed as one of the most significant legal tests of copyright law in the AI era.

What Happens Next?

The court will decide whether the publishers’ allegations warrant sanctions against OpenAI. Any ruling could influence discovery standards in AI litigation, but the underlying copyright claims will continue to be litigated regardless of the sanctions request.

The case remains ongoing, and no court has determined that OpenAI hid evidence or violated copyright law. Those issues are still subject to judicial review.