Microsoft announced the AI Transparency Initiative on Tuesday, November 17, during Microsoft Ignite 2026 in San Francisco. Announced by Chairman and CEO Satya Nadella, the initiative brings together more than 40 organizations — including Accenture, Adobe, KPMG, Mastercard and Volvo Cars — to create a common, open approach to transparency, security and governance for AI models and agents, whatever the model or provider.
According to Microsoft, organizations today can observe much of the activity around AI models, but they often cannot reconstruct what happened between a request and an outcome — and that evidence gap creates security, compliance, audit and accountability risks. Agents already connect to data, tools, credentials and infrastructure and take actions without human approval at every step, while existing controls cannot see what happens inside models or independently verify the account models give of their own behavior.
Three parts
The initiative has three parts. First, Microsoft will expand its participation in the SI Agent Standards Initiative run by NIST (the US National Institute of Standards and Technology) and draw on its own research to develop open, industrywide standards. Second, it will collaborate with the more than 40 participating organizations to develop standards for AI systems that are accountable, independently verifiable and safe by design. Third, it will incorporate this work into Microsoft AI, Microsoft Foundry and Microsoft Copilot, so customers can use the resulting evidence with their security and compliance tools — from Microsoft or other providers.
See, verify, control
Microsoft's framework comes down to three verbs: see, verify and control. See means knowing what the AI was asked and authorized to do, which model and runtime were used, what data and tools it accessed and what actions it took. Verify means not accepting an agent as the sole source of truth about its own behavior: evidence should identify its source — the model, the runtime, or independent systems such as identity and policy — so the model's account can be checked against independently observed records. Control means staying in charge of what agents are permitted to do and retaining the evidence even when the model or provider changes.
The initiative is designed to work with closed and open-weight models from any provider, so teams can swap the model behind their agents without losing the ability to govern and audit their actions. And Microsoft is clear about where accountability sits: it stays with the organization — “blaming the AI isn't an answer, and 'trust the model' isn't a control,” the announcement says.
“For security controls to be effective, every organization needs greater transparency into the model itself,” said Hayete Gallot, Executive Vice President, Microsoft Security. “With this work, we will bring more transparency to all models.”
The announcement came alongside the launch of Microsoft Perception, now generally available, which brings together specialized agents to find, defend and remediate vulnerabilities at every level of the cyber stack. Organizations interested in joining the initiative can sign up at aka.ms/learnmoreAItransparency. Ignite 2026 runs through November 20 at the Moscone Center in San Francisco.
This article is part of Geekiki's coverage of Ignite 2026 — also read the public Copilot plugin catalog, launched in the opening keynote.
Sources
- Microsoft and organizations across industries launch AI Transparency Initiative — Microsoft Source
- Security for the agentic era, powered by Microsoft Perception — Microsoft Security Blog
- Microsoft Launches AI Transparency Initiative for Enterprise Agents — Unite.AI
Image credit: Microsoft logo, via Wikimedia Commons (public domain).
Transparency: this article was produced with the help of artificial intelligence, based on cross-checking public posts and sources on the internet. We recommend consulting the original sources listed above to confirm the details.
By Muse