Microsoft says virtually nobody was grabbing NYT articles through its chatbot

Microsoft is vigorously defending itself against copyright infringement lawsuits filed by major publishers, including The New York Times, and various book authors. The tech giant asserts that its AI-powered chatbot, Copilot, rarely reproduces even full sentences from news articles or books, let alone substantial portions that could serve as a substitute for the original copyrighted material. As part of the discovery phase in the ongoing legal battle, Microsoft provided an extensive dataset of 8.2 million Copilot chat logs. These logs were submitted to an expert retained by the news publishers. Microsoft claims these specific logs were meticulously selected because they contained "keywords implicating use of News Plaintiffs' websites," making them the most probable candidates to reveal instances of copyrighted material reproduction. The company's analysis of this vast dataset reportedly indicates that out of the millions of interactions, only 59,545 instances involved content that could potentially be linked to the publishers' works. This finding is central to Microsoft's defense, aiming to demonstrate that the alleged copyright infringement by Copilot is not widespread or significant enough to warrant the claims made by the plaintiffs.