Microsoft has cited millions of Copilot conversations in its defence against a copyright lawsuit brought by The New York Times, other publishers and book authors, arguing that the records show the system rarely reproduces more than brief fragments of protected work.
The company supplied 8.2 million Copilot chat logs to an expert retained by news publishers during the legal discovery process. Microsoft said the conversations were selected because they included keywords linked to the publishers’ websites and were therefore considered the most likely to contain their material.
Its analysis found that 59,545 conversations contained at least 16 words matching news content used to train the artificial intelligence model, according to court filings. Microsoft also said an expert working for the Center for Investigative Reporting identified 51 examples of “substantial overlap” with its journalism.
In a separate analysis connected to the authors’ case, Microsoft said only 24 of the 8.2 million conversations contained responses with at least 30 matching words. Of the 212 books examined, only 10 were found to contain any matches, it claimed.
The company is using the figures to support its argument that training AI models on copyrighted material should qualify as fair use under US law. Microsoft says Copilot serves purposes different from the original books and articles, and that occasional reproduction of text “hardly undermines the transformative purpose of LLM training”.
Microsoft seeks early judgment in copyright lawsuit
The filings form part of a wider legal battle in which publishers and authors allege that Microsoft and OpenAI built their products using copyrighted works and are now competing with the organisations and individuals who created them. The claimants also accuse the systems of sometimes reproducing protected material.
The news publishers’ and authors’ cases were placed before the same judge to simplify proceedings, despite objections from the claimants. Microsoft has asked for summary judgment, which could bring the case to an end without a full trial if the judge accepts its arguments.
The New York Times, the Center for Investigative Reporting and the Authors Guild did not immediately respond to requests for comment. The Trump administration has separately filed a statement of interest in The New York Times case backing OpenAI.
