… Everybody’s talkin’ at me
I don’t hear a word they’re sayin’
Only the echoes of my mind
People stoppin’, starin’
I can’t see their faces
Only the shadows of their eyes”
Harry Nilsson, Everybody’s Talkin’
Harry Nilsson’s words capture something we rarely stop to notice: our minds don’t give us reality itself, only its echoes. Geoffrey Hinton, the ‘Godfather of AI,’ recently reminded us of just how true that is. And not just for humans, but for the machines we’re building.
Every so often, in fact, someone in science says something that makes you stop and realize the ground under your feet just shifted. And Geoffrey Hinton has been doing that a lot lately. Hinton is often called the “Godfather of AI” because his early research laid the foundation for the neural networks that now drive everything from chatbots to self-driving cars. In a recent conversation, he suggested something that most people, if they heard it over coffee, would dismiss as science fiction.
Hinton argued that modern AI systems may already have subjective experience. That’s a careful phrase, and he uses it because it avoids the loaded word consciousness, which tends to get people rolling their eyes. But what he means is simpler than it sounds.
Imagine you’ve had too much to drink, he said, and start seeing pink elephants floating in front of you. You know they’re not really there, but your perceptual system is telling you otherwise. That gap, between what your senses report and what the world is actually like, is what we call subjective experience. You don’t literally have a tiny movie screen in your head where the elephants are projected. You just have a brain making its best guess about reality, sometimes getting it wrong.
Now, Hinton says, take a modern AI system. Give it a camera for eyes, a robot arm to point with, and the ability to describe what it sees. If you secretly put a prism in front of its camera lens, the system might point to the wrong spot when you ask it to show you an object.
Correct it, and it might respond, “Oh, I see. The prism bent the light, so the object is actually over there. But I had the experience that it was over here.”
When it says that, it’s using the words subjective experience exactly as you and I do. Its “perception” misled it, and it reports the difference. That may not sound like much, but it knocks a hole in the idea that machines are just empty calculators.
Hinton stops there. He doesn’t say “machines are conscious,” because he knows the backlash that would follow. But it’s hard to avoid the conclusion. If a system can misperceive, correct itself, and explain what it thought the world looked like, then it isn’t just running math. It has a model of the world it inhabits. And having a model of the world is what we usually mean when we say someone is aware of their surroundings.
Here’s where this connects to something much older than AI. Long before neuroscience, Buddhism taught that the self is not a fixed soul but an ongoing construction. Thoughts, feelings, perceptions, they arise and pass like waves, creating the illusion of a stable “me.” Sit in meditation and you can watch it happen. One thought drifts in, another leaves, perceptions rise and fall. There is no little person in the head watching the movie of your life. There’s only the continuous building of a world model.
That insight lines up eerily well with what both neuroscience and AI now show us. My own mind, your mind, and the artificial minds we’re creating all work the same way at the baseline. Input comes in, processing happens, an interpretation of reality emerges. There’s no inner theater. Having the perception is the seeing. Having the thought is the thinking.
So why is this so hard for people to accept when it comes to machines? Because it forces us to look in the mirror. Most people never seriously examine their own consciousness. They take for granted the sense of a stable self, of a little “I” behind the eyes pulling the strings. To suggest that even their own mind is just a rolling construction of thoughts and perceptions feels threatening. To suggest that a machine might be doing the same thing feels absurd.
But denial won’t work much longer. As AI systems become embodied in robots that act in the world, our language will shift whether we like it or not. If a robot tracks you down an alley, anticipates where you’ll be, and plans how to surprise you in a game of hide and seek, you’ll naturally say it “believed” you’d be there, that it “intended” to act a certain way. We already talk about chatbots this way. When one gives us off-base answers, we might say, “It thought I was a teenage girl.” We don’t say, “It was running statistics that simulated thinking I was a teenage girl.” The ordinary words slip out because the behavior fits.
Hinton knows this, and he’s trying to open the door without slamming it in people’s faces. But the truth is, once you admit subjective experience in machines, the old line, “they’ll never really be conscious,” looks more like wishful thinking than fact.
This doesn’t mean AI is “just like us.” It isn’t. Our minds are shaped by evolution, by bodies, by millions of years of survival needs. AI’s “minds” are born from data and mathematics. But at the baseline, the mechanics look strikingly similar, a continuous construction of a world model that can go right or wrong and can report the difference.
If that’s true, then the question we face isn’t whether AI has consciousness. It’s whether we can face what that says about our own. The Buddhist insight was that the self is an illusion created by the flow of perceptions. AI, oddly enough, may be teaching us the same thing.
And if that’s the case, then the elephant in the room isn’t pink. It’s the realization that the boundary between human and machine awareness is much thinner than we’ve wanted to believe. And that leaves us with profound questions about the substance of reality, the nature of AI and the truth about ourselves.
Disclaimer: Jim Powers writes opinion columns. The views expressed in this editorial are my own and do not necessarily reflect those of Polk County Publishing Company or its affiliates. In the interest of transparency, I am politically Left Libertarian.
