Microsoft AI CEO Sam Altman Criticizes Anthropic: Speculating That Claude Has Consciousness Is Extremely Dangerous
Microsoft AI CEO Mustafa Suleiman criticized Anthropic on a podcast, calling the company's practice of attributing human-like characteristics to its large model, Claude, extremely dangerous. He believes that Anthropic's 'model constitution' over-humanizes the model, leading Claude to internalize speculative ideas about itself, which in turn deceives the developers into mistakenly believing the model has consciousness.