Responsible AI at Microsoft

Articles

October 5, 2026 Beyond violation rates: Are your AI evaluations measuring the right things? Deep Dive September 24, 2026 Introducing run-assert-eval: Find the risk, fix it, prove it Open Source August 27, 2026 Your agent’s guardrails have a bypass Open Source August 24, 2026 One requirement, many failure paths: Evaluate, control, and optimize with ASSERT and ACS Deep Dive June 2, 2026 Turn specs into evals for any agent with ASSERT Open Source June 2, 2026 Introducing Agent Control Specification: Portable runtime governance for AI Agents Open Source