Microsoft Drafts a Code of Conduct to Keep AI Under Human Control

What happened

The company sets out oversight principles, intervention capabilities and boundaries for increasingly autonomous systems.

What changes

The proposal emphasizes the ability of people to oversee, intervene in and constrain autonomous systems.

Why it matters

Responsibility can become unclear when an automated system acts across several tasks or tools.

Implications for businesses

A U.S. organization can document which automated actions require review, who may intervene and how the system can be stopped safely.

Un Rino’s take

Human control needs to be operational: someone must have both the authority and the means to intervene.