Seth, Block and Tononi: three reasons AI isn't conscious

Three of the field's most-cited consciousness researchers used a Nature commentary on 19 October to argue that today's AI systems are not conscious and probably never will be. Anil Seth of the University of Sussex, Ned Block of New York University and Giulio Tononi of the University of Wisconsin–Madison lay out three distinct reasons.

The first is that intelligence and consciousness are different things. 'Consciousness is about being and feeling; intelligence — very broadly — is all about doing, or performing functions,' the authors write. Their anatomical illustration is the cerebellum, which contains 'nearly four-fifths of the neurons in the human brain' yet whose 'absence or removal... does not affect consciousness directly.' Doing a lot does not entail feeling anything.

The second is substrate. The authors reject computational functionalism, writing that 'brains are not — or at least not just — computers that are made of meat.'

Third, they lean on two frameworks: integrated information theory, which is Tononi's own, and biological naturalism. Each treats specific wiring and specific material properties as constitutive of experience. On that reading, consciousness is 'extraordinarily unlikely, at least in the digital, silicon-based systems that underpin today's technologies.'

The authors also pre-empt the precautionary case. 'Uncertainty is not the same as knowing nothing,' they write, pushing back against the view that labs should extend moral protections to AI systems just in case. Seven researchers on our tracker had passed the piece around within days of publication, a sign it is landing in the AI-welfare debate active inside the frontier labs.