A Scientist Says There’s a Key to Being Conscious—and AI Doesn’t Have It

4 min readHere’s what you’ll learn when you read this story:While the advancement of AI has brought on fears that read like science fiction, one researcher proposes that there is no way it can be conscious without meeting certain requirements.He argues evolutionary pressures are behind developing cognition and consciousness, but the two develop independent of each other rather than consciousness coming from cognition.So long as AI is being programmed and having its needs met by humans, it will never have a need to evolve consciousness, and whether it could without those things is debatable.Far from the humanoid robots of Blade Runner and the cyberpunk genre, and despite former British Prime Minister Boris Johnson’s paranoia about machines turning sentient and plotting a massive takeover of Earth, there’s a reason today’s AI probably doesn’t possess consciousness. Digital entities whose every need is met have no reason to develop it through evolution.Consciousness is often assumed to have emerged from higher intelligence. That assumption is why the fear of AI waking up someday, taking authority into its own digital neurons and running the humans who used to run it, is so pervasive. Researcher Jeff Stibel believes the assumption is simply wrong. Stibel, who has authored books such as Wired For Thought and Breakpoint (which may or may not induce insomnia about the future of technology), argues that consciousness arose from a different set of evolutionary pressures altogether, going back billions of years in deep time. Organisms developed cognition to model a nascent and unstable planet. They developed consciousness to survive it.“At the proximate level, cognition and consciousness are different,” Stibel said in a study recently published in Frontiers in Neuroscience. “Cognition’s job is to get the world right by modeling what is out there, predicting what it will do, tracking gradients, and classifying signals. Consciousness is the management of internal conflict in a system whose own viability is at stake.”Running alongside cognition is a second capacity Stibel calls control: everything an organism does to keep itself viable. It comes in three levels. Proto-control means an organism maintains its own integrity, drawing a boundary between self and not-self, deciding what to let in and what to keep out. Meta-control arbitrates between competing internal demands, like hunger versus exhaustion, but only locally, without the whole system being involved. Settlement control is the highest form of control, resolving conflicts across every domain at once and anchoring the outcome to the organism itself.Stibel argues that only this third tier, settlement, amounts to consciousness. Proto-control and meta-control manage conflict in bacteria and thermostats, which don’t have an inner feeling of self-awareness. What changes at the level of settlement is that the resolution covers the whole organism and is tied to that organism’s own continued existence. Stibel’s further claim is that settlement does not cause consciousness or precede it. Instead, settlement is what consciousness consists of. As he puts it, “The feeling is the implementation.”The lesson Stibel draws is that cognition can grow as sophisticated as you like without control growing along with it. They are extraordinarily good at getting the world right and have no stake in the outcome. For example, with no brain or nervous system, bacteria exercise proto-cognition by following a glucose gradient. Separately, they exercise proto-control by maintaining a membrane that distinguishes their own interior from everything else. Thus, any given creature might be competent and capable of survival with cognition alone—and the same distinction between cognition and control also applies to artificial intelligence.Large language models like Gemini and ChatGPT (which humans can still beat at chess), can only pursue goals that were handed to them by humans. They’re incapable of setting their own goals, so they operate based on values we create for them. Stibel doesn’t think this is permanent, however, or that being an engineered product disqualifies AI from the potential for consciousness. “Engineering does not preclude ownership—a designed system can have owned stakes,” he writes, and “the relevant pressure need not be evolutionary in origin.” Consciousness could be possible for AI, in other words, but only in agents that must maintain themselves independently.Stibel proposes six predictions so his theory can be tested. Keeping sensory input constant while changing other variables, such as delay or conflict, will reveal whether competing demands spark more conscious engagement. Deliberately inducing failures of comparability, for example through dual-task loads, tests how conscious engagement increases or decreases relative to performance. If an AI agent hasn’t already been trained to do this, it’s unlikely to succeed on its own. Checking whether attention, memory encoding, and motor performance shift together or separately makes it possible to distinguish dynamic consciousness from mere parallel machinery. This segues into the next test, which involves disrupting the coordination of functions like memory and attention to see if consciousness was doing the work of holding them together. There should also be evidence of self-awareness, such as in some cephalopods, which may not have a centralized nervous system, but whose flexible behaviors suggest they’re aware of themselves. The sixth and final test asks whether AI can develop consciousness under pressure or whether it instead responds based on how it was programmed.“Consciousness should be possible,” Stibel theorized, “But only in agents that must maintain themselves and the conditions their operations depend on, operate under real resource budgets, use a functional self-model, and act in closed loops where consequences change future viability.”Boris Johnson can sleep easy. It doesn’t look like his sci-fi nightmare visions are about to pass through a portal and enter the realm of reality anytime soon.Elizabeth Rayne is a creature who writes. Her work has appeared in Popular Mechanics, Ars Technica, SYFY WIRE, Space.com, Live Science, Den of Geek, Forbidden Futures and Collective Tales. She lurks right outside New York City with her parrot, Lestat. When not writing, she can be found drawing, playing the piano or shapeshifting.