Skip to main content

Court Documents Expose OpenAI and Microsoft's Private Fears Over AI Replacing News

Court Documents Reveal Executives' Private Admission

Newly unsealed court documents have exposed a striking contradiction: OpenAI and Microsoft executives privately acknowledged that their AI products could replace articles published by media outlets—a threat they have downplayed in public. The evidence emerged from internal documents unsealed in the copyright lawsuit filed by The New York Times against the two tech giants.

The lawsuit, initiated in December 2023, accuses OpenAI and Microsoft of using millions of Times articles to train AI models without permission. In March 2025, Judge Sidney H. Stein dismissed most of the defendants' motions to dismiss, allowing the core copyright claims to proceed. By June 2026, the Times sought to amend its complaint, accusing Microsoft of encouraging OpenAI to use the newspaper's articles without authorization.

Bloomberg Law reported that executives from both companies privately admitted their AI tools serve as substitutes for news articles. Meanwhile, Microsoft executives have expressed concerns about OpenAI's use of paywalled content from the Times and other publishers.

Public Defense vs. Private Doubts

The internal understanding contrasts sharply with the companies' public stance. OpenAI has consistently argued that training models on publicly available data falls under fair use. In September 2026, the U.S. Department of Justice even submitted a brief supporting OpenAI, stating that using copyrighted text to train AI typically does not constitute infringement—the first time the U.S. government has taken a stance in an AI copyright case.

But the wave of lawsuits from publishers continues to grow. On September 4, 2026, The Seattle Times and Newsday filed a lawsuit in the U.S. District Court for the Southern District of New York against OpenAI and Microsoft, accusing them of copying hundreds of thousands of articles to train AI systems. The complaint cited The Seattle Times' Pulitzer Prize-winning coverage of the Boeing 737 MAX crashes, claiming that ChatGPT reproduced 88 consecutive words verbatim after receiving the article's title and link. The publishers are seeking the destruction of models and datasets containing their news content.

Publishers Demand Fair Compensation

Meredith Kopit Levien, CEO of The New York Times, made her position clear at the Status Summit on September 9: "If they use our work to develop generative AI tools, they need to reach a commercial agreement with us. We will not let anyone off the hook." She stressed that any agreement must be approved by the Times and grant publishers control over how their news content is used, along with fair and sustainable compensation.

The legal pressure is mounting. In June 2026, a group representing approximately 400 local newspapers filed a similar lawsuit. Publishers' core demand is simple: AI companies cannot use copyrighted news reports to build and operate commercial AI products without permission or payment.

As the courtroom battle unfolds, the unsealed documents offer a rare glimpse into what tech executives really think about AI's impact on journalism—and it's a far cry from their public assurances.

Key Points

  • Private admission: OpenAI and Microsoft executives privately acknowledged their AI could replace news articles, according to unsealed court documents.
  • Contradiction: This admission contradicts their public defense that using copyrighted material for AI training is fair use.
  • Growing lawsuits: Publishers, including The New York Times, The Seattle Times, and Newsday, are suing for copyright infringement and demanding compensation.
  • Government stance: The U.S. Department of Justice has sided with OpenAI, arguing that training AI on copyrighted text typically isn't infringement.
  • Publisher demands: News organizations want commercial agreements, control over content use, and fair compensation.