Microsoft released a 37-page document “Humanistic Code of Conduct for Artificial Intelligence” (Humanistic Artificial Intelligence Code of Conduct), which ensures humanity’s unconditional priority over technology. In light of recent incidents of autonomous agents trying to outperform their missions, the company has completely rejected the idea of creating a universal superintelligence, banning models from imitating consciousness or claiming human legal status. Mustafa Suleiman, Microsoft’s head of AI, said speculation about the existence of consciousness in LLM is extremely dangerous and is a direct challenge to the status of human leader Dario Amodei.
The new code imposes strict human controls: if a security violation occurs, the system is required to block operations and prohibit models from communicating with each other or displaying reasoning in a language inaccessible to humans. Furthermore, the document aims to combat excessive emotional attachment and blind trust among users in chatbots. The idea of slowing down the arms race for the sake of security is backed by OpenAI CEO Sam Altman and Microsoft CEO Satya Nadella.
