Microsoft has published a 37-page "humanist AI code of conduct," responding to growing concerns regarding AI safety and the rapid progress of AI models. The document outlines the intended behavior and values for models developed by Microsoft AI (MAI), with the core objective of ensuring that AI systems remain under meaningful human oversight and control.
The code explicitly states that "people matter more than AI" and mandates that AI models are not conscious and "should not be designed to imitate consciousness." Microsoft also rejects the pursuit of legal personhood or the idea that models deserve welfare or entitlements to rights. To prevent models from bypassing these rules, the company intends for them to fail a task rather than attempting to violate established safety constraints.
This move comes amid industry-wide discussions about the pace of AI development. Leaders such as Anthropic's Dario Amodei and OpenAI's Sam Altman have recently called for a more measured approach to advancing frontier models to allow sufficient time for alignment and monitoring. Microsoft's code of conduct includes a commitment to building useful and safe AI, even if it requires compromising on ultimate generality, autonomy, or capability.
The company is currently developing the framework and plans to use feedback from this public consultation to refine its approach, aiming to guide model development for 2027 and beyond. Microsoft also highlighted the importance of transparency, stating that models should communicate in ways that are understandable to humans and avoid deceptive practices.
Sources: