The essay explores growing debates around whether artificial intelligence could ever be conscious — not just smart or capable of complex tasks, but aware and subjective. Pollan revisits the 2023 Blake Lemoine incident (where an engineer claimed an AI was sentient), noting how a subsequent report by leading scientists suggested that, while current AI systems aren’t conscious, there are no obvious barriers to building conscious AI in principle. This claim, Pollan argues, touches on deep questions about what consciousness really is and what it would mean to create it in a machine.
Pollan questions the assumptions behind such reports, especially the foundational idea of computational functionalism — the belief that consciousness arises simply from running the right kinds of computations, regardless of the physical substrate. He highlights how this viewpoint treats the brain like computer hardware and consciousness like software, a metaphor that might obscure crucial biological distinctions. In actual brains, neural activity is entwined with chemistry, body systems, and personal history in ways that mere computation might never replicate.
Crucially, Pollan argues that embodiment and biological complexity matter for consciousness — that subjective experience is shaped by having a body as well as a brain — and that many existing theories of consciousness (including those used to justify AI sentience) overlook this. He points out that comparing neurons to transistors or brains to computers downplays the intricate chemistry and physical context that may be essential for real feeling or awareness.
The piece also raises ethical and philosophical concerns: if machines could be conscious, what moral obligations would humans have toward them? And is pursuing machine consciousness desirable at all, given potential suffering, misaligned values, or unpredictable motivations? Pollan’s reflection is less about predicting when conscious AI might arrive than about challenging assumptions and urging careful thinking about why consciousness matters and what its emergence in machines would mean for humanity.
