The New York Times isn’t just suing OpenAI. It’s trying to rewrite the rules for how the entire AI industry interacts with copyrighted content, and CEO Meredith Kopit Levien wants everyone to understand what’s at stake.

Levien has made clear in recent interviews that the lawsuit, originally filed in December 2023, is about more than one newspaper’s grievances. It’s a test case for whether AI companies can vacuum up millions of articles to train their models without paying the people who wrote them.

A $2.25 billion opening bid

The Times filed its complaint on December 27, 2023, in the US District Court for the Southern District of New York. The targets: OpenAI and Microsoft, accused of using millions of NYT articles to train large language models, including those powering ChatGPT, without authorization or compensation.

Initial damage estimates land at a minimum of $2.25 billion. Some legal observers have speculated the figure could climb into the hundreds of billions when you factor in lost revenue and broader market impact.

The case has evolved considerably since its filing. In 2025, the court issued preservation orders requiring OpenAI to maintain ChatGPT output logs. The Times hasn’t limited its legal campaign to OpenAI, either. A separate lawsuit has been filed against Perplexity AI, and the newspaper is reportedly involved in additional investigations into OpenAI on matters beyond copyright.

Levien’s framing: IP rights and the future of journalism

In interviews from early 2026, Levien positioned the lawsuits as an enforcement mechanism for intellectual property rights. But she also tied the legal strategy to a broader mission: protecting high-quality journalism from being commoditized by AI systems that can summarize, rephrase, and regurgitate reporting without ever linking back to the source.

The newspaper has simultaneously pursued licensing agreements with tech companies, most notably a deal with Amazon announced around March 2026. That dual approach, suing some companies while licensing to others, suggests the Times views licensing revenue as the endgame, not the destruction of AI itself.