Microsoft is a big house, and in this case Suleyman's criticism of Anthropic's not-so-subtle attempts to convince the world that Claude has a "soul" is entirely correct.
This is distinct from the very real safety issue of AI making dangerous information available to bad actors.