Seth, Block and Tononi: AI isn't conscious, maybe never

Three of the most-cited names in consciousness science have jointly argued, in Nature, that today's AI is not conscious and may never be on its current trajectory.

The 19 October comment by Anil Seth of the University of Sussex, Ned Block of New York University and Giulio Tononi of the University of Wisconsin–Madison lays out three reasons.

The first separates intelligence from experience. "Consciousness is about being and feeling; intelligence — very broadly — is all about doing, or performing functions," the authors write. Their anatomical case: the "absence or removal of the cerebellum, which contains nearly four-fifths of the neurons in the human brain, does not affect consciousness directly." A structure that handles most of the brain's computation can be removed and consciousness stays. Fluent chatbots mislead because training data makes them sound like us; AlphaFold, equally intelligent but non-linguistic, triggers no consciousness debate.

The second targets computational functionalism, the premise that running the right program on any hardware is enough for experience. The authors contend "brains are not — or at least not just — computers that are made of meat." The substrate matters.

The third invokes two theoretical frameworks: Tononi's integrated information theory and biological naturalism. On the IIT reading, consciousness requires components wired "in very specific, tightly interconnected ways" that today's layered, feed-forward deep-learning stacks do not share. Biological naturalism adds specific material properties, like electrochemical synapses and metabolism, that silicon lacks.

Seven of the researchers in our Who's Who tracker shared the Nature piece in the first wave of circulation.