Breaking

Microsoft Publishes Humanist AI Code of Conduct to Keep MAI Models Under Human Control

September 15, 2026 at 00:39 EDT

Foundation Models
AI Agents
Policy & Regulation

Microsoft AI has published a draft Humanist AI Code of Conduct for its MAI models, setting out a people-first framework that says advanced AI should remain subordinate to humans and subject to meaningful oversight. The document is not yet being used to train models; Microsoft says it is open for a six-week public consultation and will be revised toward the end of 2026 before guiding model development in 2027 and beyond.

Sep. 14, 2026 · Microsoft AI

Human Control Becomes The Rule

Microsoft AI published a draft Humanist AI Code of Conduct for MAI models, saying advanced AI should remain subordinate to people and fail rather than violate safety rules.

pages in the draft code

parts in the framework

weeks of public consultation

2027

target for model guidance

Five-Part Governance Stack

5 partsCode structure

The code organizes Microsoft AI's future model rules into mission, safety constraints, uncertainty guidelines, defaults, and concluding appendices.

What supporters like

A clear hierarchy says models should fail a task rather than violate safety constraints.

What skeptics watch

The draft is not yet used for training, and its real test is whether Microsoft accepts slower progress when safety requires it.

The Control Loop

Human goal→Code rules→Model action→Fail if unsafe

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.