Seth, Block, Tononi: AI isn't conscious, likely never will be

Three leading consciousness researchers argue in Nature that today's AI systems are not conscious and that the dominant research path probably will not get them there. 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 reasons the field, in their view, keeps mistaking fluency for mind.

The first is a definitional reset. "Consciousness is about what it feels like to be something, while intelligence is about what a system can do," the authors write. They point to the cerebellum, which contains four-fifths of the neurons in the human brain and performs sophisticated computation yet plays no direct role in conscious experience. Language models, they argue, exploit the opposite confusion. "LLMs are mirrors that reflect an image of our collective digitized past," the trio writes. "We gaze into them and all too readily confuse the reflections with our own conscious minds."

The second reason rejects computational functionalism, the premise that running the right program on any hardware is enough for experience. Substrate, they argue, is doing real work. The third reason leans on specific scientific theories: integrated information theory, which ties consciousness to a tightly interconnected architecture absent from the layered, feed-forward structure of contemporary deep-learning models, and biological naturalism, which grounds consciousness in material properties of living brains that silicon does not share.

The essay lands in a year when AI labs themselves have started probing their own models for signs of inner life. The authors note that Anthropic published research earlier in 2026 claiming "computational signatures of a form of consciousness in Claude," work that faced substantial criticism, including from Seth himself. Seven researchers tracked in our Who's Who directory shared the Nature piece the same week.

Their closing position is deliberately flat: current systems should be treated as "smart tools, not smart people," and moral concern should stay with beings whose capacity to suffer is not in doubt.