The Long View

Understand the news. Get on with your day.

In context

Court documents show Microsoft and OpenAI worried about news use

First published 18 September 2026, 13:09 UTC.

What happened

Newly unsealed court documents in The New York Times's copyright lawsuit against Microsoft and OpenAI showed internal concern within both companies over their use of millions of news articles to train AI systems, the Times reported. The Washington Post reported that one of the unsealed documents was headlined by a quote calling the practice a possible 'largest theft of labor' in history.

Why it matters

The underlying case could help decide whether AI companies must license news content or can continue training on it without payment, a question with major financial stakes for publishers and AI firms. Publishers point to the newly unsealed documents as evidence the companies understood the risks to journalism, while Microsoft and OpenAI can still argue in court that their use of the material is protected as fair use.

How we got here

The New York Times sued Microsoft and OpenAI in December 2023, alleging the companies trained AI models on its content without authorization and that their products could reproduce copyrighted articles in ways that competed with its journalism. Separately, the Authors Guild and other authors sued OpenAI in September 2023 over similar training practices involving books. Those cases were later consolidated before a single federal judge in New York, who in 2025 allowed most of the claims to proceed toward trial. The newly unsealed material reported this week adds internal documents to that consolidated litigation.

How we got here, dated

  1. 2023NYT sues Microsoft and OpenAI. The New York Times filed a copyright infringement lawsuit against Microsoft and OpenAI, alleging unauthorized use of its articles to train AI models. Source
  2. 2023Authors Guild sues OpenAI. The Authors Guild and 17 authors filed a separate class action against OpenAI over training on their books, one of several related 'Author Actions' later consolidated in the same court. Source
  3. 2025Judge allows most claims to proceed. Judge Sidney Stein denied several OpenAI motions to dismiss parts of the Times's case, letting most claims move forward while dismissing some narrower claims. Source
  4. 2025Ruling on AI-generated summaries. Stein denied OpenAI's bid to dismiss claims that ChatGPT-generated summaries could themselves infringe copyright, finding a jury could see them as substantially similar to originals. Source
  5. 2026Court documents unsealed. Newly unsealed filings in the consolidated case showed internal concern at Microsoft and OpenAI over using news articles to train AI systems, drawing wide coverage. Source

A useful comparison

Authors Guild v. Google (the 'Google Books' case). A content rights holder sued a major tech company over mass, unauthorized use of copyrighted works to build a tech product, with the case turning on whether the use was fair.

Where the comparison breaks down: Courts found Google Books only displayed limited snippets and did not substitute for the original books, a fact pattern different from the NYT dispute, which centers on internal admissions that AI outputs can directly substitute for journalism.

What remains unclear

  • The full unsealed court record has not been independently reviewed here; reporting relies on documents quoted in press coverage.
  • How the newly unsealed material will affect pending summary judgment motions in the consolidated case is not yet known.
  • Some figures circulating in secondary coverage about the case have not been verified against primary court filings.

What to watch

Watch for how the court handles the newly unsealed material as it rules on pending summary judgment motions in the consolidated copyright case.

Sources and evidence

Read the original reporting at the links above. Our analysis can be wrong and may change as evidence develops.

Related background

Court documents show Microsoft and OpenAI worried about news use | The Long View