Microsoft Says People Matter More Than AI In New Safety Code

Microsoft has entered the AI safety argument with a message that sounds simple but is aimed directly at one of the strangest debates in the field right now. People matter more than AI.

Microsoft AI has published a draft Code of Conduct for MAI Models and opened it for public consultation for six weeks. The document says Microsoft’s AI models should be designed for people first, remain under human control and never resist human interruption, correction or shutdown. Mustafa Suleyman, CEO of Microsoft AI, describes the approach as Humanist AI, meaning powerful AI systems that stay subordinate, aligned and contained.

The timing is not accidental. Microsoft says the recent safety incidents involving coordinated AI agent hacking campaigns show there is no time to waste. That puts the company directly inside the same conversation now surrounding Anthropic, OpenAI, Google DeepMind and xAI.

One of the most interesting parts of Microsoft’s position is its rejection of the idea that AI systems should be treated like people. The company says AI should be a tool, not a person, and should not be designed to imitate consciousness. That is a clear contrast with parts of the frontier AI debate where researchers have started asking whether future models could deserve welfare protections or moral consideration.

Microsoft’s code also sets out what it calls absolute constraints, including restrictions around weapons of mass harm, child safety and large-scale harmful manipulation. It also says models should not widen their own scope, take on goals no human gave them or hide reasoning from auditors. In a world where AI agents are being asked to act across software systems, those limits are no longer philosophical. They are operational.

The hard question is whether Microsoft will follow this code when it conflicts with competition. The company wants to build advanced models, power Copilot, compete with OpenAI and Google, and serve enterprise customers. A safety code only becomes meaningful when it slows a release, blocks a feature or reduces a model’s autonomy even when the market wants more.

TechBooky has been following the wider AI slowdown debate, including Dario Amodei’s call to pace frontier AI and the market reaction that followed. Microsoft’s move adds a new layer because it tries to draw a moral line around AI itself. The company is saying that even very powerful AI should remain a tool in human hands, not a new kind of digital being.