{"aiVersion":"1","content":{"id":"cmu698fub0004777ywgbi3550","slug":"when-news-became-training-data-what-the-unsealed-f-20260918","title":"When News Became Training Data: What the Unsealed Filings Say","level":"MEDIUM","publishedAt":"2026-09-18T01:03:07.372Z"},"topic":{"slug":"when-news-became-training-data-what-the-unsealed-f-20260918","category":"technology"},"article":{"paragraphs":["The unsealed filings in the New York civil case have revealed internal messages and documents. The filings say OpenAI's mid-training datasets held 91,692 copies of works from the Times, the Daily News and the Center for Investigative Reporting. They also describe a Common Crawl-derived set that included more than two million nytimes.com documents and note that Microsoft and OpenAI moved training data through projects like Project Mango and Project Taxi.","News plaintiffs said the materials make clear that the products can be substitutes for original reporting, while OpenAI head Nick Turley warned the tools posed an existential threat and wrote that the products are largely substitutive, period. The filings say OpenAI researchers discussed plans to bypass paywalls and that some training data had copyright notices removed. Microsoft presentations described a doom loop that would hurt publishers and the web, and Microsoft data has been cited showing that Copilot answers cut click-through rates to nytimes.com by as much as 93%. Project Mango is said to contain at least 160,903 unique works, and company leaders including Satya Nadella have been deposed about licensing paywalled content."],"wordCount":183,"readTime":2},"vocabulary":[{"word":"dataset","example":"The researchers built a dataset of news articles.","phonetic":"/ˈdeɪ.tə.sɛt/","definition":"A collection of data used together for analysis or training."},{"word":"substitute","example":"An AI answer can act as a substitute for a full article.","phonetic":"/ˈsʌb.stɪ.tjuːt/","definition":"Something that replaces another thing."},{"word":"paywall","example":"Many sites use a paywall for premium articles.","phonetic":"/ˈpeɪ.wɔːl/","definition":"A system that limits access to online content unless you pay."},{"word":"infringement","example":"The lawsuit alleges copyright infringement.","phonetic":"/ɪnˈfrɪndʒ.mənt/","definition":"A breach of someone else's legal rights, such as copyright."}],"quiz":[{"answer":"91,692","question":"How many copies did OpenAI's mid-training datasets hold from the publishers named?"},{"answer":"Project Mango","question":"Name one project used to move training data between Microsoft and OpenAI."},{"answer":"Nick Turley","question":"Which OpenAI executive warned that the tools posed an existential threat and wrote that the products are largely substitutive, period?"}],"discussion":[{"question":"Do you use short summaries from search or AI instead of full articles? Why or why not?"},{"question":"Have you noticed fewer visits to news sites when using answer engines? What changed?"},{"question":"What would you do if a site you like added a strict paywall?"},{"question":"Do you trust AI summaries of news? Why or why not?"}]}