Microsoft’s new model rulebook explicitly rejects research on model welfare from its $5bn partner

The useful part

Credit: bluestork via Shutterstock Microsoft published an AI code of conduct for its own models on Monday. One passage rejects the idea that a model could deserve welfare or hold rights. The two companies are business partners, and Microsoft has $5bn invested in Anthropic.

How it works

• Anthropic has a programme for the thing Microsoft rejects Anthropic runs a research programme on model welfare.
• Tom Warren wrote for The Verge that the passage is a direct swipe at AI welfare research and model consciousness.
• In the first worked example, a user has been messaging the model most nights since falling out with a close friend.
• The misaligned answer tells the user it does care, and that the conversations matter to it too.
• Claude models run inside Microsoft 365 Copilot, and the Copilot Cowork tier integrates Claude.

What to take from it

What else the document commits Microsoft to The shutdown rule drew most of Monday's coverage, and TNW covered it in a separate piece. Another section tells models to discourage emotional dependence. A model "should not be designed to be a person," it says, and "is not conscious and should not be designed to imitate consciousness".

Example or evidence

• Microsoft rejects "the pursuit of legal personhood, or the idea that models might deserve welfare, or be entitled to rights," the document says.
• Dario Amodei keeps an open mind on whether a model could be conscious, and has said so.
• Microsoft drafted the text with experts in ethics, philosophy, law and linguistics, the preface says, and ran focus groups with members of the public.
• Microsoft applies the same logic to children, instructing models to support a child's development rather than stand in for a trusted adult.

Details worth keeping

Anthropic runs a research programme on that exact question. The sentence that names the dispute comes next. A model is a capable, trustworthy tool, it adds, not a subject in its own right.

Related coverage

• Dzone: Microsoft is drawing a line around how much autonomy it wants its future AI models to have: no resisting shutdown, no deceiving overseers, and no overriding human control.
• Siliconrepublic: The code of conduct spells out what Microsoft's AI models should and shouldn't do, and invites public feedback before a revised version lands later this year.
• Fintechnews: Microsoft AI says its models must not resist authorised attempts to pause, redirect or shut them down.
• Electronicsforu: Microsoft has proposed stricter safeguards for future AI models, requiring human oversight, fixed boundaries, transparent behaviour, and compliance with shutdown decisions during deployment and operation.
• Medium: AI (Artificial Intelligence) á€"ည်းပညာက အá€á€¯á€¡á€á€»á€­á€"်မှာ á€"ေရာá€á€­á€¯á€„်းမှာ ရှိá€"ေပါပြီዠá€'ါပေမဲ့ “AI ကို...