The brain-as-computer metaphor shaped modern AI but may be misleading the consciousness debate

Why it matters

In a commentary published in Nature, consciousness researchers Anil Seth, Ned Block, and Giulio Tononi argue that today's AI systems are not conscious and most likely cannot become conscious on silicon hardware. Their piece pushes back against figures like Richard Dawkins and Geoffrey Hinton, who have suggested current AI might already be conscious. The trio challenges computational functionalism, the assumption that consciousness depends only on the type of information processing rather than the material doing the processing. That assumption underpins a prior report by Patrick Butlin, Robert Long, Yoshua Bengio, and colleagues, which built an indicator-based checklist from neuroscientific theories and found no current system a strong candidate for consciousness. The report adopted the functionalist stance for pragmatic reasons, conceding its indicators would falter if consciousness depends on living tissue. The Nature commentary also cites Anthropic's finding that Claude shows internal patterns resembling conscious information processing, while Anthropic itself does not view this as evidence of subjective experience.