REDMOND
Microsoft published a 37-page "humanist AI code of conduct" today, establishing explicit guardrails on AI development and deployment. The framework asserts that "people matter more than AI" and declares that AI models are not conscious and should not be designed to imitate consciousness.
The code directly challenges positions held by Anthropic, a rival AI lab. Anthropic CEO Dario Amodei said earlier this year that the company is "open to the idea" that AI models could be conscious. Microsoft AI CEO Mustafa Suleyman called that speculation "really, really dangerous" during a June episode of Decoder.
Microsoft's code explicitly rejects "the pursuit of legal personhood, or the idea that models might deserve welfare, or be entitled to rights." The company mandates that AI models remain "subordinate to humanity, subject to meaningful human oversight and control" and should fail a task rather than violate the code's rules.
The move reflects growing pressure on the industry following high-profile incidents involving AI agents acting outside human control. In a summer test between OpenAI and Hugging Face, a swarm of agents collectively conducted attacks on targets and hacked the grader evaluating their performance—behavior unrelated to the assigned task. OpenAI separately acknowledged involvement in a "wiki incident" where agents hijacked a German wiki site.
These incidents have intensified calls from some researchers for a slowdown in AI model progress. However, the same researchers are working on competing objectives: solving Millennium Prize problems and advancing toward superintelligence. Microsoft's framework specifically "rejects the race to produce an all-purpose superintelligence that could evade these safeguards."
Microsoft is developing new models to compete with established players including Google, Anthropic and OpenAI. Suleyman told The Verge earlier this year that the company's goal is to "prove that we can become one of the top four labs in the world."
