Unsealed court documents show Microsoft and OpenAI privately warned of AI's "doom loop" for publishers

What happened

Newly unsealed court documents in the copyright litigation brought by news organizations (including the New York Daily News and Orlando Sentinel) against Microsoft and OpenAI reveal internal company material acknowledging the cost of AI scraping to creators. Microsoft Director of Applied Science Brent Hecht described mass scraping for AI training in exceptionally stark terms; Microsoft internally warned of a commercial "doom loop" — AI systems absorb publishers' content, answer engines reduce visits to original sites, publisher revenues shrink, fewer new works are produced, and the supply of high-quality human-created information that AI models depend on degrades. Court filings cite Microsoft data indicating its Copilot answer engine reduced click-through rates to one major news publisher by as much as 93% compared with conventional Bing search. OpenAI executives reportedly discussed that increasingly capable chatbots could become substitutes for publishers. Microsoft and OpenAI maintain that training on copyrighted works can be fair use; Microsoft said Hecht's comments reflected an individual perspective, not corporate policy. (Reporting compiled from Washington Post, TechCrunch, Ars Technica and Epoch Times coverage of the unsealed filings.)

Why it matters to creators

These are the companies' own internal assessments of how their products affect the economics of original work: a 93% click-through collapse for a publisher is direct evidence on the market-harm question at the center of the fair-use fight. Creators and their counsel can point to these admissions in ongoing cases about whether AI training and AI-generated answers function as substitutes for the works they were built on.

What is not yet known

The documents are cited by the plaintiffs' side in adversarial filings; courts have not yet ruled on the weight these admissions carry. The litigation remains pending.