Nadella Slams AI Double Standard: Trains on Others' Data but Bans Distillation

The Decoder · rss · 2026-07-13

Microsoft CEO Satya Nadella publicly criticized AI labs like OpenAI and Anthropic for a "reverse information paradox." He pointed out that while these companies scrape public data to train models under the guise of "fair use" and learn from user interactions, their terms explicitly forbid others from distilling their own models. Nadella urged enterprises to build and control their own learning infrastructure, which conveniently aligns with the core cloud and AI services Microsoft is currently pitching to customers.

Original post →

More from Companies & People

Companies & People channel →