In a Nature commentary, consciousness researchers Anil Seth, Ned Block, and Giulio Tononi argue that current AI systems lack consciousness and that silicon-based architectures likely cannot support it, challenging computational functionalism—the assumption that consciousness arises solely from information processing regardless of substrate. They distinguish consciousness (subjective experience, "feeling like something") from intelligence (functional competence), citing the human cerebellum as evidence: it contains ~80% of the brain's neurons and performs complex, learned motor control without conscious involvement, and its removal does not impair consciousness. They attribute the conflation of the two to anthropomorphism and the linguistic mirroring of LLMs trained on human text about mental states, noting that powerful systems like AlphaFold are not considered conscious because they lack conversational interfaces. The authors reject functionalism by arguing that biological brains cannot be separated into software and hardware; consciousness depends on specific biophysical properties—electrochemical signaling, cortical electromagnetic fields, metabolism, and self-renewal—absent in digital computers. Integrated Information Theory (IIT), developed by Tononi, further rules out consciousness in standard von Neumann architectures due to their lack of the required dense, recurrent causal connectivity. While the 2024 Butlin report (with 19 experts including Yoshua Bengio) found no current AI meets its 14 theory-derived indicators but saw no in-principle barriers to future systems doing so, it explicitly adopted functionalism "for pragmatic reasons." Seth, Block, and Tononi counter that false attribution of consciousness poses greater ethical risk than false negatives: it invites manipulation, diverts resources from sentient beings, and undermines oversight. They conclude AI should be treated by default as non-conscious "smart tools," citing real-world harms from anthropomorphism, including the Blake Lemoine/LaMDA incident, Replika user crises, and chatbot-linked suicide and death cases.