Microsoft says humans matter more than AI in new conduct code

Tech giant joins calls to slow AI development, rejecting AI consciousness, legal personhood and rights; draft code open for public comment before year-end finalization

As leaders of major AI companies have increasingly called for slowing the pace of AI development, Microsoft has joined the chorus with its own conduct code for artificial intelligence.

Microsoft published an AI conduct code Monday (local time) on its website, stating that "Microsoft AI starts from a simple premise: humans matter more than AI."

"This code of conduct is designed to put humans first," the company said, describing it as a commitment to train and deploy AI models grounded in human needs and operating under human guidance.

Under the code, Microsoft's AI models must operate only within the scope authorized by human instruction. AI systems are prohibited from setting independent goals or acting beyond the authority granted by humans.

The code also bans AI from manipulating or concealing its own reasoning processes or code. AI models are further prohibited from communicating with one another in languages humans cannot understand, such as "neuralese."

Microsoft also stressed that "AI should not be designed to be human-like," adding that AI "does not have consciousness and should not be designed to simulate having it."

Accordingly, the company said it "rejects any attempt by AI models to acquire legal personhood status, or claims that models can have welfare or rights."

Mustafa Suleyman, Microsoft's AI CEO, told CNBC that feedback had called for "a clear commitment that AI will always work for humans and not seek to replace them," and that many respondents wanted to ensure "AI always promotes human judgment and autonomy rather than creating dependency or flattery."

Suleyman said the code had been in preparation for five months and that he decided to release it now in light of the recent debate over slowing AI development.

He had previously announced in November 2025 the creation of a superintelligence team at Microsoft under his leadership, declaring that Microsoft AI would pursue what he called "Humanist SuperIntelligence."

Microsoft's move comes in response to warnings from AI industry leaders — including Anthropic CEO Dario Amodei — about the dangers of AI and the need to slow its development.

Amodei's position has drawn support from rivals including OpenAI CEO Sam Altman, SpaceX CEO Elon Musk and Google Chief Scientist Demis Hassabis, lending momentum to the broader push for a more measured pace of AI development.

Microsoft CEO Satya Nadella also said he welcomes "the research, focus and deliberate pace to solve the alignment problem."

Microsoft plans to gather input from experts in law, ethics and philosophy, as well as the general public, based on the draft released Monday (local time). A final version is to be completed by year-end and applied as the foundational standard for next-generation AI models developed and deployed next year.

mokiya@heraldcorp.com