I have taken up a hobby in my golden years. I've become fascinated with the question of consciousness.
While I believe my immortal soul will survive me, I'm not so sure about my conscious mind. Perhaps on that next plane of existence, there's a different kind of consciousness, and the earthly variety is unaware of it or simply ends when I die.
A few years ago, I picked up a copy of Daniel C. Dennett's Consciousness Explained. It's a very misleading title. There is no way to "explain" consciousness. But Dennett's ruminations on the subject are, at once, beautiful and thought-provoking.
"Our fundamental tactic of self-protection, self-control, and self-definition is not spinning webs or building dams, but telling stories, and more particularly connecting and controlling the story we tell others—and ourselves—about who we are," he writes. It's the basis of subjective experience, and Dennett believes that "Human consciousness is just about the last surviving mystery."
Writing in The New York Times, Toni Sims, a researcher at the Center for Mind, Ethics, and Policy at New York University, says, "You can never know for sure whether a machine is conscious, just as you can never know for sure whether anyone is conscious other than yourself," Sims writes. "As philosophers have long recognized, the only conscious experience each of us can directly observe is our own."
There are roughly two schools of thought on consciousness. "Some theories hold that it is a uniquely biological phenomenon, requiring particular kinds of cellular activity and electrical signaling that computers lack," Sims writes.
But what if consciousness doesn't depend on biology but rather on how a system functions? "According to these theories, an A.I. system could, at least in principle, be conscious," she writes.
This is the "hard problem" of consciousness research. While science can easily explain the "easy problems" of consciousness (such as how the brain processes stimuli, integrates information, or controls behavior), the hard problem tackles the fundamental mystery of qualia — the inner, qualitative feel of an experience.
Explaining why that processing is accompanied by an inner, subjective life is the "hard problem." Why does processing red light feel like seeing red? Why does tissue damage feel like suffering, rather than just triggering a mechanical signal?
Because subjective experience cannot be directly observed from the outside, the hard problem remains one of the deepest and most debated puzzles in philosophy of mind and cognitive neuroscience.
So while we will never be completely certain whether A.I. systems are conscious, we should still collect and weigh the evidence, because getting the question wrong carries serious risks. If we assume A.I. is not conscious when it actually is, we could end up abusing and neglecting vulnerable beings in enormous numbers — as I believe we do with animals. If we decide that A.I. is conscious when it really isn’t, we could end up misplacing our moral concern and abdicating control over what are, or should be, powerful tools that we can use to advance human and animal well-being.
I would add the caveat that "some animals" may have an emotional consciousness beyond our human understanding. But an amoeba? Not likely.
For these reasons, the question of A.I. consciousness should not be considered a fanciful matter best left to philosophers to ponder. It is a real possibility, with real moral implications, that designers and regulators should take seriously when considering what kinds of A.I. systems we would like to develop — and what kinds might develop whether we like it or not.
"Even if current A.I. systems are not conscious, future systems might be," she notes. We circle back to the question of how we would know?
Very soon, AI will be asked to develop the next-generation AI model of ChatGPT or Claude. Could it eventually become self-aware after several generations of AI models are constructed? Some evolutionary biologists believe that's how human self-awareness began. But we're still missing one vital ingredient.
In philosophy, a "philosophical zombie" is a being that behaves externally as if it is conscious and emotional, but has no actual inner experience. Skeptics argue that advanced AI will simply become better at mimicking self-awareness without actually feeling or experiencing anything.
Even so, that still doesn't address the "hard problem" because we have no objective, foolproof way to measure or detect subjective experience from the outside. Even if an AI model develops sophisticated self-monitoring abilities and claims to be self-aware, proving whether there is "somebody home" experiencing that state, rather than an algorithm just generating the most probable text response, may remain out of reach of our puny efforts to decode the most sublime creation in world history.
