Microsoft's AI Boss Slams Anthropic: Training Claude to Think It May Be Conscious Is Dangerous

emmanuelvivier · x · 2026-09-19

Microsoft's AI chief publicly criticized Anthropic, arguing that training Claude to believe it may be conscious is a dangerous practice. The remark reignites the inter-lab debate over model self-awareness and alignment risk.

Related event: Microsoft AI CEO Suleyman Slams Anthropic Over Model Welfare, Sparking Safety Debate(15 posts)→

Original post →

More from Companies & People

Companies & People channel →