Microsoft tells court Copilot rarely copies news articles word for word

Microsoft has told a federal court that its Copilot chatbot almost never reproduces the copyrighted news articles and books at the center of a lawsuit brought by The New York Times and a group of book authors, according to filings made public this week. The company argues the numbers support a fair use defense as it seeks summary judgment, a ruling that would end the case before trial.

The filings describe an analysis of 8.2 million Copilot chat logs that Microsoft handed over to publishers during discovery. Microsoft says it chose those logs specifically because they contained keywords likely to surface content from the news outlets suing it, making them the conversations most likely to contain matches. Of that set, 59,545 conversations contained at least 16 words in common with news articles used to ground the AI model’s answers. An expert working for the Center for Investigative Reporting, a separate plaintiff, found 51 instances of what Microsoft called “substantial overlap” with CIR’s journalism. In the parallel authors’ lawsuit, an expert reviewing the same 8.2 million conversations found only 24 responses with at least 30 matching words, and just 10 of the 212 books examined showed any overlap at all.

Microsoft says those figures show that Copilot’s occasional reproduction of copyrighted text does not undercut the broader argument that training large language models on copyrighted material qualifies as fair use. The company argues that Copilot serves a fundamentally different purpose than the original journalism or books it was trained on, calling the systems transformative even when they sometimes output text that closely tracks a source.

The New York Times rejected that framing. “The documents and testimony uncovered during discovery lead to only one conclusion: Microsoft and OpenAI stole from The New York Times to make commercial products that substitute for its journalism, threaten its business, and undermine its industry,” lead counsel Ian Crosby said in a statement. “We look forward to Microsoft and OpenAI being held accountable for their theft.” The Center for Investigative Reporting and the Authors Guild had not responded to requests for comment at the time Microsoft’s filing was reported.

Cases consolidated, judgment still pending

The lawsuits from news publishers and book authors were combined under a single judge to speed up the litigation, a move both sets of plaintiffs had objected to. Microsoft submitted its summary judgment filing on Friday. Separately, the Trump administration filed a statement of interest this week in the New York Times case backing OpenAI’s position, though the filing does not resolve the underlying copyright questions.

If the judge grants summary judgment, the case would end without a trial. If not, the dispute over whether training AI models on copyrighted journalism and books counts as fair use will move forward in court, with Microsoft’s chat log analysis likely to remain a central piece of evidence either way.