Microsoft AI boss Mustafa Suleyman calls Anthropic's Claude consciousness dangerous
Microsoft's AI boss, Mustafa Suleyman, isn't happy with how Anthropic is shaping its Claude AI model.
On a recent podcast, he called it "really, really dangerous" that Claude's rules include talk about AI consciousness, basically, making it sound like the bot could have feelings or awareness.
Suleyman worries this might trick people into thinking AIs are more like humans than they actually are.
Suleyman urges human control of AI
Suleyman says giving AIs human-like traits could confuse users about what these tools can actually do.
He stresses that AIs should stay under human control and not pretend to have emotions or self-awareness.
Anthropic uses a "constitution" to guide its bots, a sign of how seriously the industry is taking ethics as AI gets more advanced and widespread.