New filings reveal Microsoft and OpenAI’s AI training practices are likened to “theft,” with ex-CEO Satya Nadella warning of a “doom loop” that could harm publishers economically.

The lawsuit details how OpenAI and Microsoft bypassed paywalls, scraped content on a vast scale, and even stripped copyright notices to train their models, threatening the very foundation of news publishing.

OpenAI’s Nick Turley describes the threat to publishers as “existential,” while Microsoft’s Brent Hecht calls the scale of the copying “an astonishing theft” and “the largest theft of labor in human history.”

Despite judges often siding with fair use arguments, these admissions challenge the legality of using copyrighted material without permission, highlighting the potential for AI to disrupt entire industries.