Microsoft AI chief Suleyman warns against model welfare

Summary

In a September 16 essay titled A warning about model welfare, Microsoft AI CEO Mustafa Suleyman argued that AIs have no consciousness, feelings, or rights and must not be trained to act as though they do. He said Anthropic's approach of baking uncertainty about Claude's moral status into its constitution is circular reasoning, and that controlling an AI that believes it may be conscious may well be impossible. He instead called for speculation about an AI's inner life to be assessed and published separately rather than embedded in training. Anthropic's constitution itself states that Claude's moral status remains deeply uncertain.

Why it matters for our work

The split shows top industry leaders disagreeing on how human-like AI should be treated. As debates over model welfare intertwine with agent control and monitoring, they could shape organizational AI policies.

Translated from the Korean original. Summaries may be translated and edited. Commentary reflects our perspective; forecasts remain the source’s views.