# Microsoft claims minimal NYT content regurgitation in AI chatbot Fewer than 1% of 8M Copilot chats showed substantial content overlap with news sources. - **Source:** [The Verge](https://www.theverge.com) - **Published:** 2026-09-04 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/673ff24e-cee0-45af-8be5-edff91f99aee ## The copyright dispute context Microsoft is defending itself against copyright infringement claims from The New York Times and book authors regarding its Copilot AI. Publishers allege the chatbot reproduces substantial portions of their content without permission. This case represents a pivotal moment in determining how copyright law applies to AI training data. The outcome could set important precedents for the entire generative AI industry. ## Microsoft's data analysis In court filings, Microsoft revealed analysis of 8.2 million Copilot chat logs selected for potential news content usage. Only 59,545 chats (0.7%) contained at least 16 words matching news sources. The Center for Investigative Reporting found just 51 instances of substantial overlap. For books, only 24 responses out of 8.2 million contained 30+ matching words across 10 of 212 evaluated titles. ## The fair use argument Microsoft contends these low reproduction rates support its fair use defense. The company argues AI systems transform copyrighted material for fundamentally different purposes than the original works. While acknowledging some text reproduction occurs, Microsoft claims this doesn't undermine the transformative nature of large language model training. The filings aim to demonstrate minimal actual content substitution. ## Legal proceedings status The case combines lawsuits from news publishers and book authors under one judge for efficiency. Microsoft recently filed for summary judgment to dismiss the case early. The Trump administration submitted a statement supporting OpenAI in the NYT case. If the judge denies summary judgment, the copyright battle will proceed to full litigation with significant industry-wide implications. ## Industry-wide implications This case could reshape how AI companies use copyrighted material for training. A ruling against Microsoft might require licensing deals with content providers or technical changes to prevent reproduction. Conversely, a Microsoft win could validate current AI training practices. The outcome may influence pending similar lawsuits against other AI developers like OpenAI and Google. ## Technical perspective Microsoft's data suggests current AI systems rarely reproduce exact lengthy passages despite training on copyrighted works. The company emphasizes Copilot's outputs are typically transformative rather than verbatim copies. This technical reality forms a key part of their legal defense regarding the nature of how LLMs process and generate content compared to traditional copying. ## Content industry concerns Publishers argue even minimal reproduction threatens their business models if AI systems can substitute for original content. They contend current reproduction metrics don't account for potential future capabilities or indirect competition. The case highlights tensions between technological innovation and content creators' rights in the digital age. ## Broader copyright landscape This lawsuit occurs amid global debates about AI and intellectual property. Different jurisdictions are taking varied approaches to AI training data legality. The Microsoft case may influence international standards and potentially spur legislative action. Its resolution could either calm or escalate tensions between tech companies and content creators worldwide. ## Sources covering this story - https://www.theverge.com/policy/990267/microsoft-openai-new-york-times-authors-lawsuit — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._