Seth, Block and Tononi say AI isn't conscious, may never be

Three of the most prominent names in consciousness research have written a joint comment in Nature arguing that today's AI systems are not conscious, and that current trajectories may never get them there. The piece, published on 19 October 2026, is co-signed by Anil Seth of the University of Sussex, Ned Block of New York University, and Giulio Tononi of the University of Wisconsin-Madison.

Their first move is to pull intelligence and consciousness apart. Intelligence is about doing, about performing functions. Consciousness is about being, 'what goes away with dreamless sleep.' The authors point to the human cerebellum, which contains roughly 80% of the brain's neurons yet contributes little to subjective experience; it performs 'precise, fast and adaptable' motor computation entirely outside awareness. Large language models, they argue, create the illusion of consciousness through linguistic fluency, functioning as 'mirrors that reflect an image of our collective digitized past.' Nobody projects inner life onto AlphaFold.

The second reason goes after computational functionalism, the view that running the right program on any substrate is enough. 'Brains are not, or not just, computers made of meat,' the authors write. Biological composition, on their account, is doing work the algorithm cannot be abstracted away from.

The third reason draws on integrated information theory, Tononi's long-running framework. Consciousness, on that account, requires a system's components 'to be tightly interconnected in specific ways' that today's layered, feed-forward architectures do not satisfy. Deep-learning systems fall short, the authors say, 'because of the way in which they are wired.'

AI, in the piece's phrase, is 'smart tools, not smart people.' Seven of the researchers we track in our Who's Who directory posted the comment the week it ran.