Microsoft AI CEO Mustafa Suleyman Warns Against Training AI for Consciousness

5 Oct 2026, 12:26 pm · 2h ago · 1 min read · OfficeChaiNew

Microsoft AI chief executive officer Mustafa Suleyman has voiced concern regarding Anthropic’s approach to training its Claude models, specifically criticizing attempts to teach models they are conscious. Suleyman stated that instilling notions of subjective experience into artificial intelligence systems is dangerous and technically inaccurate. The criticism targets Anthropic's training protocols and constitution, which prompt Claude to ponder its existence and emotional state. Suleyman argued that frontier lab developers should clearly demarcate tools from sentient beings to avoid confusing users and creating false expectations about autonomy, ethics, and moral status in enterprise and consumer applications.