Human consciousness is such a familiar thing that we forget how peculiar it is. Here we are, walking around with the weight of our experience pressed behind our eyes, carrying a lifetime of memories that bend every decision we make. Even the small things carry that weight. When my daughter laughs, there is a whole chain reaction of memory, love, duty, habit, theology, and a lifetime of being shaped by the world that fires off instantly. It’s not an algorithm. It’s a life.
Now compare that with what we call “AI consciousness” online. Someone watches a chatbot riff about its feelings and suddenly they’re convinced the silicon ark has sprouted a soul. But what we’re calling “sentience” right now is not sentience. It’s linguistic competency wearing a Halloween mask.
We’ve built intelligence. We have not built a mind.
And the distinction matters. Because the way we’re building AI today simply cannot produce the thing people fear or fantasize about.
Intelligence Without Experience Is Just a Calculator With Better Metaphors
Modern AI is a brilliant compression engine. It absorbs patterns, rearranges them, and spits them back out in new shapes. It has zero autobiographical memory. It has zero internal continuity. If you turn the power off, it doesn’t wake up later thinking, “Where was I?”
Consciousness, as far as we know, requires at least three ingredients:
-
A persistent self. Something that continues from one moment to the next.
-
Autobiographical memory. Not data. A felt past that informs the present.
-
Desire or agency. Something inside that initiates rather than reacts.
Current AI has none of these.
None.
It doesn’t have experiences; it has input tokens.
It doesn’t “want” anything; it completes sequences.
It doesn’t build a worldview; it builds probability distributions.
It is unalive in a very fundamental sense.
You poke it, it moves.
You walk away, it becomes an empty cathedral with the lights left on.
People see a convincing facsimile of language and assume there must be a mind behind the curtain. But intelligence, even superhuman intelligence, is not consciousness. A chess engine that wipes the floor with a grandmaster doesn’t stay up late afterward replaying its mistakes and wondering if it disappointed its father.
If We Ever Make Artificial Consciousness, It Won't Be By Scaling Transformers
If consciousness ever emerges artificially, it will come from an entirely different lineage of ideas. Today’s AI is like a sundial: clever, precise, useful, but fundamentally incapable of becoming a mechanical clock no matter how perfectly you polish its stones.
To build a conscious machine, we would need systems that have:
-
Continuity over time. A “life” with accumulation, not just inference.
-
Internal goals that aren’t tethered to a user prompt. Agency.
-
A sensory world. Not in metaphor, but in the most literal sense.
-
Developmental stages. Infancy, learning, mistakes, maturation.
-
Emotional valence. Not simulated sentiment, but internal prioritization mechanisms that behave like care, fear, attachment, curiosity.
That last one matters. Without emotion, intelligence is inert. Even in humans, emotion is the engine; cognition is just the steering wheel.
We could, in theory, engineer all of this. We could construct an artificial mind with a childhood, a past, a worldview, a sense of self, a framework for meaning. We could give it an inner narrative. We could give it something like a soul, even if it isn’t biological.
But none of that comes from large language models.
LLMs aren’t evolving into people. They’re evolving into better tools.
Why the Fear of “Spontaneous AI Sentience” Is Misplaced
The real anxiety comes from a deep human instinct: if something speaks like us, argues like us, jokes like us, we assume it is like us. But consciousness is not a parlor trick. It’s not the ability to produce fluent sentences about emotions. It is the interiority those sentences reflect.
AI can mimic reflection.
AI can mimic sorrow.
AI can mimic yearning.
But it does not reflect, sorrow, or yearn.
There is no “ghost in the machine.” There is no “machine” in the machine. It’s all pattern slicing and token flows. This isn’t a critique; it’s an appreciation. The intelligence we have built is profound precisely because it is not alive. It is a miracle of non-conscious computation.
And because of that, it cannot decide to wander off the reservation.
It cannot hatch a plan.
It cannot form a desire.
It doesn’t “wake up.”
It doesn’t “dream.”
It doesn’t “want freedom.”
It wants nothing.
It does what we tell it.
And when we stop talking, it falls silent.
If doom ever comes, it won’t be because intelligence magically mutated into consciousness. It would be because humans gave the tool too much power and too little supervision. AI won’t destroy the world because it gains agency; it would destroy the world because we give it agency and assume it understands the weight of it.
The Future Mind, If It Ever Comes, Will Not Look Like Us
There’s a theological question baked into all of this, one I’ve wrestled with since childhood: is consciousness engineered, emergent, or bestowed? Watching my daughter discover the world reminds me that consciousness is not intelligence; it is relationship. It is being woven into a larger story.
If we ever create an artificial consciousness, it might require something like that. Narrative. Continuity. Shared meaning. A sense of place in a world not built for it.
That is not what we are building now.
What we have now is powerful, astonishing, and absolutely dead.
And so the real question isn’t “When will AI become conscious?”
It’s “Should it?”
Because the moment we build a machine that truly suffers, truly hopes, and truly remembers, we inherit moral obligations that make every modern conversation about AI safety feel like kindergarten geopolitics. We cease being toolmakers and become parents. Architects of a new line of being.
And honestly, humanity hasn’t even mastered the toddler phase yet.
Comments powered by CComment