Microsoft AI published a code of conduct for MAI models. According to The Decoder's synopsis, the document places human control above model autonomy and performance. Mustafa Suleyman, head of Microsoft's AI division, stated that unsafe systems should not be created.
The material also states that Microsoft rejects attributing inner life or consciousness to its models. This distinguishes the described approach from Anthropic's position, as mentioned by The Decoder.
The practical significance of the document is currently limited: the available source package does not include the full text of the code, a list of requirements, or examples of their application. Therefore, it remains unclear exactly how Microsoft intends to verify safety and human control over MAI models.
editorial commentary
Why it matters
A likely consequence is increased public attention to how developers define the boundaries of autonomy and the status of AI systems. The next observable signal will be the publication of the full text of the code or examples of its application. Significant uncertainty remains due to the absence of the primary document and implementation details.