We keep talking about AI as if it introduced the black box.
A system speaks. It answers. It persuades. It appears to reason. But its interior remains inaccessible. We see input, output, behaviour, weights, activations, traces. We do not see experience. We do not see understanding in any direct sense. We do not get inside. So we call it a black box.
Fair enough.
But the more I think about it, the less convinced I am that the black box arrived with artificial intelligence at all.
AI did not create the condition.
It made it newly difficult to ignore.
The black box was already in other minds, long before it was in machines. You have never accessed another person’s interior directly. You infer it from voice, expression, gesture, language, pattern, delay, contradiction, warmth, withdrawal. You read the surface and construct a model. Sometimes the model is close. Sometimes it is disastrously wrong. Either way, the interior itself remains unavailable.
This is not a niche philosophical problem. It is the ordinary structure of relation.
But even that formulation now feels incomplete to me.
Because the black box was not only in other minds.
It was already in the self.
And not only in the self as “mind,” but in the body.
That is where this gets more interesting.
We often speak as though the body were the transparent bit. The hardware. The visible surface. The measurable mechanism through which some hidden interior expresses itself. The body, under that view, is the accessible layer; mind is the inaccessible one.
But that division is too neat.
Maurice Merleau-Ponty saw this much earlier than most contemporary AI discourse. In Phenomenology of Perception, the body is not a dumb vessel for a ghostly interior, nor a machine waiting to be objectively described from outside. It is the lived body: the medium through which world, self, and relation first appear at all. You do not stand outside it and consult it like a manual. You inhabit it. You move through it. You are oriented by it.
And yet, precisely because of this, it never becomes fully transparent.
The body is intimate without being exhaustively knowable. It is immediate without being fully given. It is present, functional, structuring — but never wholly disclosed. Try to observe it directly and it slips from lived medium into object. Try to grasp it completely and you end up with anatomy, measurement, representation. Useful, yes. Complete, no.
The hand that touches is also touchable. The eye that sees cannot quite turn around and catch itself in the act of seeing. The body is not a window. It is a relay.
That word matters to me.
A relay is not pure origin and not pure destination. It is a crossing point. Something passes through it, but not unchanged. Signal is stabilised, transformed, delayed, interpreted, carried. A relay does not erase mediation. It formalises it.
That feels much closer to what the body actually is.
Not transparent container.
Not sealed essence.
Relay.
The more technological life becomes, the harder this is to deny. Donna Haraway, in A Cyborg Manifesto, treated boundaries not as sacred natural facts but as unstable constructions: human and machine, animal and tool, body and system, all less cleanly divided than we pretend. Her cyborg is not just a science-fiction image. It is a way of naming the already-crossed condition we inhabit: patched, mediated, hybrid, interfaced.
We are not approaching some future moment in which the body becomes technological.
We are already there.
Some people feel that truth more explicitly than others.
In my own case, hearing has never felt like a simple transparent sense. It has often felt closer to routing. Filtering. Compression. Partial reconstruction. Signal passing through systems, technical and biological, before arriving as anything like usable perception. Cochlear implant, hearing aids, crossed wires — not as dramatic identity markers, but as quiet reminders that perception is not some clean untouched stream from world to self.
It never was.
A device does not suddenly make perception artificial. It makes visible a condition that was already there: that access is mediated, that sense is assembled, that relation depends on partial coupling rather than pure immediacy. The interface does not destroy the body’s truth. It reveals its structure.
This is why I increasingly think the current obsession with AI opacity is both justified and slightly misplaced.
Justified, because we are clearly building systems whose behaviour is legible while their interior status remains unresolved. We can interpret outputs, audit performance, map architecture, even attach moral and legal pressure to system actions. But the deeper question — what the system is, what kind of presence or absence lives behind the interface, whether “behind” is even the right metaphor — remains open.
Misplaced, because it encourages the fantasy that opacity is an anomaly introduced by machines.
It isn’t.
The black box was already in the mind.
And before that, it was already in the body.
The body is where the problem becomes harder to sentimentalise. It is difficult to maintain fantasies of total access when your own orientation to the world is already crossed by sensation, prosthesis, memory, expectation, training, language, fatigue, mood, and inference. You do not receive reality raw. You do not receive yourself raw. What you get is structured arrival. Stabilised trace. Coupling. Enough contact to act, never enough to abolish ambiguity.
This also changes how I think about transparency.
Transparency gets treated as an unquestioned good in contemporary culture. More visibility. More data. More disclosure. More extraction. More legibility. We act as though the interior, if it exists, ought eventually to be made fully available — to institutions, to platforms, to partners, to diagnostics, to systems of measurement. The inability to do this is framed as failure. Friction. Imperfection. Bug.
But what if opacity is not merely a lack of access?
What if it is part of what relation is?
Not because mystery is romantic. Not because all opacity is noble. Some forms of obscurity are simply domination wearing a fog machine. But there is another kind: the irreducible distance that prevents relation from collapsing into total possession. The fact that the other is not fully mine to decode. The fact that self-knowledge never reaches zero distance. The fact that bodies, signals, and minds meet through approximation rather than conquest.
That is not failure.
That is dignity.
Body as relay is the hinge here: between philosophy and object, between phenomenology and interface, between posthumanism and design, between the self as lived and the self as indexed. It also quietly bridges things I had not fully connected before — hearing technology, crossed wires, Process Zine’s long-running concern with signal and noise, and SUBSTRATE.HOST’s artefacts as mute relays that do not open but point.
Maybe that is why the current AI moment feels so charged. Not because machines have suddenly introduced an alien black box into the world, but because they have forced us to confront the fact that black boxes were always the rule. The older fantasy — that selves, minds, and bodies were ultimately transparent if only we had enough data — is becoming harder to maintain.
The machine becomes unsettling because it reveals a structure we recognise too well.
Input. Output. Inference. Projection. Relay.
No direct access.
The black box was never only in AI.
It was already in the body.
Continuation: The Boundary Is the Self
Sequence note:
These essays form a connected sequence on black boxes, selfhood, mediation, and ontological limits. Together they move from the problem of AI systems we cannot define, to the self’s failure of full self-access, to the body as relay rather than transparent vessel, and then to the self as a moving boundary condition rather than a hidden interior. The sequence continues in The Moving Edge, a bridge-object that gathers these consequences into a compressed boundary metaphysics.
Read in sequence:
Leave a Reply