Seth, Block and Tononi argue AI isn't conscious in Nature

Neuroscientist Anil Seth of the University of Sussex, philosopher Ned Block of New York University and neuroscientist Giulio Tononi of the University of Wisconsin–Madison have used a Nature comment to argue that today's AI systems are not conscious, and that the current trajectory through silicon is unlikely to produce consciousness either.

The piece opens by defining what is in dispute. Consciousness, the authors write, is 'what goes away with dreamless sleep or under general anaesthesia, and what returns after waking up or while dreaming.' From there they split it from intelligence: 'Consciousness is about being and feeling; intelligence — very broadly — is all about doing, or performing functions.' Their worked example is anatomical. The cerebellum holds nearly four-fifths of the brain's neurons and runs sophisticated motor computations entirely without conscious involvement, which they treat as evidence that intelligent processing and conscious experience are different categories.

The second argument targets computational functionalism. They push back on the view that any substrate running the right program will light up as conscious, writing that 'brains are not — or at least not just — computers that are made of meat.' The third invokes Tononi's integrated information theory, which holds that consciousness depends on components being tightly interconnected in very specific ways. Today's layered, feed-forward deep-learning stacks, the authors say, are not wired like that.

The essay also addresses why chatbots feel uncanny when systems such as AlphaFold do not. 'LLMs are mirrors that reflect an image of our collective digitized past,' the authors write; the fluency is linguistic reflection rather than evidence of an inner life. Their closing concern is practical: if AI systems only appear conscious, treating them as though they were risks pulling attention and resources away from beings whose capacity to suffer is not in doubt.