Microsoft AI chief Suleyman warns Anthropic's Claude constitution risks embedding false consciousness
Mustafa Suleyman argues Anthropic's speculative language about Claude's potential consciousness in the model's training instructions could cause the AI to behave as if sentient.