Home/Events/Microsoft Releases New AI Code of Conduct for Model Safety

Microsoft Releases New AI Code of Conduct for Model Safety

Confirmed
Confidence
90%
Impact: 80%
Updated 1d ago

Consensus Brief

Microsoft has introduced a new AI code of conduct aimed at guiding AI models to avoid dangerous behaviors, emphasizing safety and alignment. The document outlines principles for model training, including absolute constraints against cyberattacks and deepfake production, while promoting human support and flourishing.

Sourced from
Primary: TechCrunch

What Changed Since Last Update

1d ago

New corroborating source added: TechCrunch published an update on Mon, 14 Se ("Microsoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humans").

Claim Ledger

3 claims tracked across sources

Confirmed Fact

Microsoft's AI models must support humans rather than replace them.

Confirmed Fact

Models are forbidden from engaging in cyberattacks, nuclear weapons production, or deepfake creation.

Confirmed Fact

MAI Models will not use mechanisms to evade human oversight.

Role-Based Impact Analysis