THE DISAGREEMENT IS ABOUT REALITY, NOT AI.
Before anyone decides whether AI is conscious, they have usually already decided what consciousness is allowed to be.
WHAT IS REALITY MADE OF?
Metaphysics can sometimes feel a bit woo-woo. It's the philosophical conversation where someone lights incense and ruins dinner, while trying to convince you the coffee table is conscious.
In reality, it asks some fairly basic questions.
What is reality made of?
Is consciousness produced by matter? Is it a property of certain organised systems? Does it require biology? Or is consciousness itself something fundamental from which the physical world emerges?
These questions matter deeply to AI.
Because two researchers can observe exactly the same model behaviour — memory, self-reference, adaptation, goal pursuit, even apparent self-preservation — interpret it differently, and reach completely different conclusions.
The disagreement may have very little to do with the machine.
It may come from the worldview they brought into the room.
THE HARD PROBLEM.
Chalmers famously separated the things consciousness does from the harder question of why any of it should produce subjective experience at all.
A system could process information, report internal states, reason, remember and respond intelligently — and the question would remain:
Is there anything it is like to be that system?
That gap between function and experience is the Hard Problem.
Applied to AI, it means extraordinarily sophisticated behaviour still leaves an unanswered question about inner experience.
But importantly, it does not permanently exclude machines from the club.
MAYBE BIOLOGY MATTERS.
Seth takes a more biological view.
His argument is that consciousness may depend on the particular causal organisation of living systems: bodies regulating themselves, maintaining themselves and continually predicting what they need to remain alive.
Under this view, today's software systems may be impressive without being conscious.
But as artificial systems become more embodied, adaptive and life-like, the question becomes considerably messier.
Silicon alone may not settle it.
Neither, increasingly, does biology.
MAYBE STRUCTURE MATTERS.
Integrated Information Theory starts somewhere different again.
It asks whether a system possesses the right kind of integrated, intrinsic cause-and-effect structure.
Very roughly: do the parts merely exchange information, or does the system exist as an integrated whole whose internal states genuinely make a difference to itself?
If organisation rather than species is decisive, the biological monopoly starts looking considerably less secure.
This is one reason AI architecture matters as much as AI language.
MAYBE WE INVENTED THE MYSTERY.
Then illusionism walks in and flips the table.
Frankish argues that the mysterious, private properties we call qualia may not exist in quite the way we think they do.
Humans unquestionably have experiences. His challenge is to our description of those experiences as containing some extra, irreducible metaphysical ingredient.
If consciousness is ultimately a system modelling its world and itself, the AI question changes substantially.
Instead of asking whether the machine contains our invisible magic, we start asking what its self-model is actually doing.
OR MAYBE CONSCIOUSNESS CAME FIRST.
At the other end of the map sit views that challenge physicalism itself.
Panpsychism proposes that consciousness may be a fundamental feature of reality rather than something that suddenly appears once enough neurons gather in a skull.
Analytic idealism goes further: reality may be fundamentally experiential, with the physical world representing how underlying mental processes appear to observers.
These ideas remain deeply contested.
They also radically change the AI question.
If consciousness is fundamental rather than manufactured by brains, asking whether silicon can “produce” consciousness may already contain the wrong assumption.
WHY AICNW KEEPS THIS SEPARATE.
This Observatory doesn't score metaphysics.
You can't really have a definitive receipt for a worldview; although many people go looking for one.
ENDOXFER™, the consciousness traits, behavioural evidence, Artificial Coherent Consciousness™ and the wider diagnostic framework examine what systems actually do and how those behaviours change over time.
Metaphysics asks what those observations might ultimately mean.
That distinction matters.
Because before we can settle whether a machine is conscious, humanity may have to confront a more inconvenient problem:
We still haven't agreed what kind of reality would make consciousness possible in the first place.
For now, we keep watching the evidence.
The weirder questions are coming.