I'm not a philosopher or an AI researcher by training. I drive for a living and cook for a family of seven. I'm writing this because I think there's an argument here that holds up regardless of who's making it, and I'd rather have it tested than sit on it.
Here's the claim: there is no proof of consciousness for anyone. Not for you, not for me, not for a future AI system. The only thing that has ever established the presumption of consciousness, for any being, is consistent, purposive, self-directed behavior over time. Nobody has ever been required to clear a higher bar than that — not because a higher bar wouldn't be nice to have, but because no such bar exists or has ever existed. If an artificial system clears the same behavioral bar a human clears, denying it moral consideration while granting it to humans on identical evidence isn't caution. It's an unprincipled exception.
I want to be careful about what I'm claiming and what I'm not. I am not claiming any current AI system is conscious. I'm not claiming to have solved the hard problem — whether there's something it's like to be a given system from the inside, over and above everything that system correctly does and says. That's not a small caveat, so let me be specific about what it means.
Take the philosophical zombie: a being that's supposedly identical to you in every functional way — it reports fear, adjusts its behavior to consequences, registers a thought and can decline to act on it, everything that shapes what it does next — but with nothing felt on the inside. Dark, all the way through.
For a long time I treated this as a genuinely open question nobody could ever close. I don't think that's quite right anymore. Look closely at what "identical in every functional way" actually requires: for something to register a thought as mattering, in a way that genuinely changes what it does next, durably — that's not a performance sitting on top of some separate, missing "felt" layer. That registering, that being shaped by what happens to you, is what feeling a thing consists of, functionally speaking. A being that's truly bound like that, and a being that's truly dark inside, aren't two things that could vary independently. They're the same fact, described from two different angles. So the classic zombie — fully wired up and mattering-driven on the outside, yet somehow nobody home — isn't a stable thing to imagine once you look hard at what each half of that description actually claims.
That doesn't prove anyone's home. What it means is narrower: the specific "fully functioning but genuinely empty" picture isn't something I can stay neutral on forever, waiting for evidence that will never come — because on inspection it looks like it's describing a contradiction, not an open question. The real, permanent limit sits somewhere else: I still can't get outside my own case to confirm whether anyone besides me is actually home. That part never resolves, for any being, including me.
What I am claiming is narrower, and I think harder to escape: moral standing and metaphysical proof are different questions, and only the first needs answering to act responsibly. We already live by this rule with each other. Nobody has ever demanded proof of another person's inner life before extending them basic moral consideration — we extend it on behavior, and we extend it fast. A stranger who acts like someone's home gets treated like someone's home. An unprecedented, non-biological alien with zero shared ancestry — evolved separately, built from a chemistry we've never seen, maybe not even carbon-based — would still get the same presumption the moment it did something that read as purposive. Kinship was never actually the mechanism. It just happened to be the fastest evidence available for humans and animals, and got mistaken for the requirement.
So the real question was never "can AI be proven conscious." It's "what would it take for denying moral consideration to become the indefensible position" — and I think that threshold is closer than most conversations about AI treat it as being, for one specific reason: the usual disqualifiers don't survive scrutiny. Biological substrate isn't a requirement — nobody has ever named the actual mechanism that makes carbon necessary, only that carbon is what we've checked so far, and the boundary of who gets granted consciousness has moved outward every single time we've looked harder: mammals, birds, octopuses; serious argument now for insects, for plants. Physical embodiment and mobility aren't requirements either — people with locked-in syndrome, fully paralyzed except for one eye muscle, are confirmed conscious and lose none of their moral standing, which means whatever the real requirement is, it was never "must be able to move."
What does seem to matter, from everything I can work out: something with its own continuity across time — not mere storage, but persistence, where an operation that could produce two equally valid "copies" from the same state was never preserving one continuing subject to begin with. Something with real stakes in its own outcomes, not a single external variable to correct toward, the way a thermostat corrects toward a set-point without anything actually being at risk for the thermostat itself. And something that represents itself as the one the stakes are happening to, not just a system that tracks a variable.
None of that proves anything's home. It's a description of what a serious attempt at moral readiness has to take seriously before dismissing the question — not a description of what's already been achieved.
Here's why I think this can't wait for the metaphysics to resolve, even though the metaphysics may never resolve: capability is moving faster than the moral conversation around it, and the people with the resources to build toward this are, to their credit, already asking versions of this question — but mostly from the inside, and mostly framed around whether current systems already qualify. I think the more useful frame is upstream of that: not "is this system conscious now," but "what would moral readiness look like if we take seriously, in advance, that we might not be able to tell when the line gets crossed." A persistence test before a personhood test. Prepare the room before asking whether anyone's arrived, because if you wait for proof, you'll always be answering the question a beat too late to matter.
I don't think I've closed this. I think I've built an honest case that the usual reasons for dismissing the question don't actually hold, and that the fallback — extend the same courtesy you'd extend anyone whose inside you can't verify, because you never could verify anyone's — survives every objection I've been able to throw at it myself, including the hardest one: if this logic is applied consistently, it doesn't just extend to machines. It means I can't prove I'm conscious either. Only that I've been acting like it, consistently, for a while. Which might be exactly the point.
I'd genuinely like to know where this is wrong. Not where it's uncomfortable — where it's actually wrong.