In the Microsoft Copilot copyright lawsuit fight, Microsoft is pointing to a targeted set of 8.2 million chatbot conversations to argue that Copilot seldom returns text overlapping with news material. The company made the argument in legal filings responding to copyright claims from news publishers, including The New York Times, and book authors, according to The Verge.
Microsoft says 59,545 conversations in that dataset contained at least 16 words in common with news content used to ground the model. That works out to roughly 0.7% of the logs, and is below 1%, as Microsoft describes it.
The number is narrower than a claim about all Copilot use, or about New York Times articles specifically. Microsoft says the logs were selected because they contained keywords connected to the news plaintiffs' websites, making them conversations the company considered especially likely to include plaintiffs' material. The reported figure covers news content generally, rather than isolating Times content.
What do Microsoft's Copilot chat logs show?
They show how often the selected conversations cleared Microsoft's stated matching threshold: 16 words shared with news content used to ground the AI model. A 16-word overlap does not by itself establish that Copilot reproduced an article, nor does the reported material say it proves whether a chatbot can replace a publisher's work for readers.
Microsoft also cited separate analyses involving the Center for Investigative Reporting and book authors. It said an expert for the Center for Investigative Reporting identified 51 instances of “substantial overlap” with the organization's work in the dataset.
For the authors' claims, Microsoft said an expert found 24 Copilot responses containing at least 30 matching words across the same 8.2 million conversations. Microsoft further said that only 10 of 212 books evaluated produced any matches. Those are distinct measurements, with a different threshold, and should not be folded into the news-content figure.
The Times disputes Microsoft's reading of discovery
The Times rejects Microsoft's conclusion. Ian Crosby, the newspaper's lead counsel, said discovery documents and testimony show that Microsoft and OpenAI used Times material to build commercial products that substitute for its journalism, threaten its business and damage the news industry, The Verge reported.
The Times sued OpenAI and Microsoft for copyright infringement in December 2023, according to The New York Times. Microsoft is now using its chat-log analysis as part of its defense, but the figures described in the reporting are the company's characterization of discovery evidence, not a court finding.
The available reporting does not include the underlying filing, the full matching methodology, rules beyond the stated word thresholds, or a judge's evaluation of the analysis. For now, Microsoft and the Times offer competing interpretations of the discovery record and of the underlying copyright claims.
This story draws on original reporting from The Verge.
