What happened
Microsoft announced (July 27, 2026) the External Red Team Alliance (EXTRA), providing unrestricted gifts to 18 university labs across six continents plus a distributed operational network of external red-team specialists to test frontier AI systems for safety/security failure modes.
Why it matters
Formalizes crowdsourced, global AI red-teaming beyond a single vendor's internal team, addressing multilingual/regional blind spots in frontier model safety testing — a notable governance/testing capability expansion, though not itself a deployable product.
Applicability
AI safety/red-team leads at labs and enterprises evaluating third-party AI red-teaming programs; academic security researchers seeking funded engagement.