Newly unredacted filings from The New York Times’ copyright lawsuit against OpenAI and Microsoft reveal that senior executives privately described AI companies’ use of scraped content as theft and a serious threat to publishers. The disclosures may undermine OpenAI’s argument that training AI models on copyrighted material is protected as fair use, particularly because the companies’ products could replace visits to original news websites.

The filings allege that content was gathered through mass scraping, including bypassing paywalls and removing copyright notices, although much of the material comes from The Times’ legal brief and the underlying exhibits remain sealed. Microsoft data reportedly showed Copilot reducing click-throughs to The Times’ website by up to 93% compared with traditional Bing search, while Microsoft CEO Satya Nadella said paywalled material should require licensing and that he would have demanded retraining if he had known it was used. OpenAI executives also acknowledged that chatbots could pose an “existential threat” to publishers and become increasingly substitutive.