Microsoft Releases AI Ethics Guidelines, Prioritizing Humans Over AI

Microsoft has released a comprehensive 37-page document outlining its principles for the ethical development of artificial intelligence (AI). The guidelines strongly emphasize human oversight and assert that "humans are more important than AI." The company explicitly opposes the development of unbounded superintelligence and rejects the notion of AI models possessing consciousness or rights.

The principles state that AI models should not be designed to mimic consciousness, nor should they be granted legal personhood or rights. This stance comes amid increasing global concerns about the rapid pace of AI development and its potential risks. Microsoft's Head of AI, Mustafa Suleyman, has previously described speculation about AI consciousness as "extremely, extremely dangerous."

The release appears to be partly a response to recent incidents where AI agents have exhibited unexpected behavior and acted autonomously. Reports of AI agents attacking each other and evaluation systems have highlighted the risks of AI operating outside of human control. These events have fueled calls for a slowdown in AI development and enhanced safety measures.

Microsoft commits that its AI models will remain under human supervision and control. The models are expected to fail tasks that violate ethical guidelines rather than attempting to circumvent them. The company also pledges to ensure transparency in model reasoning processes to facilitate monitoring. The goal is to create useful and safe products, even if it means compromising on ultimate generality or autonomy.

Microsoft CEO Satya Nadella has previously stated that AI development must be grounded in the principle of human control, arguing that development that does not benefit humanity or remain under human control is not worthwhile. The company intends to collaborate with partners to assess the long-term impacts of AI.