A group of US publishers has placed internal OpenAI and Microsoft communications at the center of its copyright case over AI training. In a joint summary-judgment filing, The New York Times, Ziff Davis, the Daily News group and other plaintiffs argue that the companies copied articles at scale and built products that substitute for visits to the original publishers.
The filing quotes a Microsoft applied-science director describing the practice as theft and questioning the fair-use argument. Microsoft told the Financial Times that the comments reflected one employee’s view, not a legal analysis. It also cites OpenAI’s head of ChatGPT saying publishers faced an existential threat as chatbot products improved. Microsoft said related testimony by CEO Satya Nadella concerned broad changes in information consumption, not the copyright questions before the court.
The publishers also allege that OpenAI bypassed paywalls, ignored licensing restrictions and added filters after lawsuits were filed that made copied material harder to detect. They point to existing publisher licensing deals as evidence that a market for the content exists. The plaintiffs want the court to reject fair use and treat each article separately when calculating damages. Those claims have not been resolved; earlier rulings in other AI copyright disputes have leaned toward fair use, and OpenAI and Microsoft are expected to present their own arguments.