Unredacted court filings allege Microsoft and OpenAI called AI training data scraping 'theft' and undermined publishers

Newly unsealed court filings in *The New York Times*’ three-year-old copyright lawsuit against OpenAI and Microsoft include internal statements from executives and researchers that describe AI training data practices as 'theft' and an 'existential threat' to publishers.

A Microsoft executive described the companies’ AI training practices as 'theft,' while OpenAI leadership characterized its models as posing an 'existential threat' to journalists and publishers whose work trained the systems, according to the unredacted material.

The filings allege that OpenAI and Microsoft obtained and used copyrighted content by bypassing paywalls, building training datasets via mass scraping, and deliberately stripping copyright notices from the data before it reached the models.

Microsoft’s Copilot 'answer engine' reportedly caused click-through rates for *The New York Times*’ domain to drop as much as 93% compared to traditional Bing search, according to an internal Microsoft presentation cited in the filing.

An internal Microsoft document warned of a 'doom loop' in which generative AI could 'hurt the performance of our models and the entire web' by undermining the economic foundations of content suppliers.

Microsoft CEO Satya Nadella testified that 'anything that is paywalled should be licensed by anyone who wants to use it…for grounding or training,' and said he would have required OpenAI to retrain its models if he had known it scraped paywalled information.

OpenAI’s head of ChatGPT, Nick Turley, wrote internally that publishers face an 'existential threat' from products like the chatbot, which he described as 'largely substitutive' and likely to become more so as models improve.

The filings also reveal that OpenAI’s mid-training datasets contained more than 91,692 copies of works published by *The New York Times*, the *Daily News*, and the Center for Investigative Reporting, with a Common Crawl-derived dataset including more than 2 million documents from nytimes.com alone.