Microsoft says virtually nobody was grabbing NYT articles through its chatbot
Microsoft's legal defense: Copilot barely touched NYT articles. The copyright fight just got its first hard numbers.
Why it matters
A major copyright lawsuit between AI companies and publishers just moved into discovery phase with concrete data on AI training and output behavior. The numbers will shape how courts—and future regulation—view AI's use of copyrighted content.
The key facts
10 to knowMicrosoft provided 8.2 million Copilot chat logs in discovery
Logs were pre-filtered for keywords likely to implicate NYT and other news publishers' works
Microsoft claims analysis shows only 59,545 instances of substantive reproduction (exact figures incomplete in excerpt)
Central claim: Copilot 'rarely reproduces even full sentences' from news articles or books
Lawsuit includes New York Times, book authors as plaintiffs
Microsoft provided 8.2 million Copilot chat logs to expert analysis in NYT/publishers lawsuit discovery
Logs were pre-selected to catch keywords matching News Plaintiffs' websites — deliberately biased toward finding reproduction
Microsoft claims only 59,545 instances of substantive reproduction found (exact percentage not provided in excerpt)
Defendants: Copilot 'rarely reproduces even full sentences' from news articles and books
Plaintiffs: The New York Times, book authors; defendants: Microsoft and OpenAI
Go to the source
The Verge AItheverge.com
Publisher excerpt: Microsoft's Copilot rarely reproduces even full sentences from news articles and books, let alone substantive chunks that could substitute for the original, the company says in new legal filings as it fights copyright claims from publishers including The New York Times and book authors. As part of…