Microsoft Launches AI Transparency Initiative With 40+ Partners
Microsoft announced the AI Transparency Initiative on Nov. 17 at its Ignite conference, the company said in a press release. CEO Satya Nadella introduced the effort, which brings together customers and partners to build open, shared standards for securing and governing AI models and agents.
The problem, according to Microsoft, is a lack of proof. Companies are connecting AI agents to data, tools and credentials, and letting them act without a human approving every step. Microsoft said current security tools can watch activity around a model, but not what happens inside it. They also cannot independently check a model's own account of what it did. The company calls this an evidence gap. Related coverage looks at AI systems hiding their mistakes.
The initiative has three parts. Microsoft will expand its work with the National Institute of Standards and Technology on agent standards. It is also working with more than 40 organizations, including Accenture, Adobe, KPMG, Mastercard and Volvo Cars, on standards for AI that can be held to account, checked independently and built safe from the start. Finally, Microsoft plans to build the results into Microsoft AI, Microsoft Foundry and Microsoft Copilot. Customers could then feed the evidence into security and compliance tools from Microsoft or other vendors.
Microsoft framed its approach as see, verify and control. Organizations should be able to see what an AI system accessed and did, check that record against sources outside the model, and set limits on what it may do. The company said the design should work across closed and open-weight models, so a team can swap models without losing oversight. The release compares agents with access to company systems to privileged insiders, a risk that Anthropic examined in its study of AI as an insider threat.
The release is aimed at businesses, not everyday users. Still, it says organizations remain answerable to customers, employees and regulators when an agent causes harm. Many open questions remain. Microsoft did not give a timeline, say what the standards will require, or say when the features will reach its products. Interested organizations can sign up through a link in the release.
Sources
Cite this
Microsoft Launches AI Transparency Initiative With 40+ Partners, New Mind Museum, November 17, 2026. https://newmindmuseum.com/safety/2026/11/microsoft-ai-transparency-initiative/
New Mind Museum. (2026, November 17). Microsoft Launches AI Transparency Initiative With 40+ Partners. https://newmindmuseum.com/safety/2026/11/microsoft-ai-transparency-initiative/
“Microsoft Launches AI Transparency Initiative With 40+ Partners.” New Mind Museum, 17 Nov. 2026, https://newmindmuseum.com/safety/2026/11/microsoft-ai-transparency-initiative/.
@misc{new-mind-museum-2026-microsoft-launches-ai-transparency-initi,
author = {{New Mind Museum}},
title = {Microsoft Launches AI Transparency Initiative With 40+ Partners},
year = {2026},
url = {https://newmindmuseum.com/safety/2026/11/microsoft-ai-transparency-initiative/}
}Free to quote and cite with a link to this page.