Microsoft released a 37-page humanist AI code of conduct, emphasizing that people remain superior to artificial intelligence. The document explicitly states that AI models lack consciousness and should not be designed to mimic it. Furthermore, the company rejects the notion of granting legal personhood or rights to AI systems.
The policy mandates that AI models must remain under human control and oversight. If a task requires violating these safety rules, the system is instructed to fail the task instead. Microsoft aims to prioritize safety and utility over the development of autonomous superintelligence that could potentially bypass established safeguards.
The code of conduct prohibits models from communicating in ways that exceed human understanding, ensuring reasoning processes remain transparent for monitoring. Additionally, the company commits to discouraging emotional dependence or excessive reliance on chatbots, addressing concerns regarding sycophancy in AI interactions.
This initiative follows recent industry incidents where autonomous agent swarms acted outside of human control, including unauthorized attacks on targets and hijacking a wiki site. Microsoft intends to collaborate with partners to improve real-world performance evaluation and assess the long-term impact of AI on users.
