The Deliberateness of “Artificial”

Why did they call it artificial intelligence?

Not machine intelligence.

Not computational intelligence.

Not simply intelligence.

Artificial intelligence.

I keep coming back to that word.

And I don't mean “deliberateness” necessarily as in some secret meeting where developers sat around a table and said, “One day humans will blame the machines, so let's call it artificial and cover ourselves.”

I mean something more interesting.

What if the choice of the word artificial—whether consciously or unconsciously—anticipated something fundamental about the way human cognition works?

Because humans have a peculiar difficulty with something that behaves like a thing without actually being that thing.

Our brains respond to what they perceive.

A movie is an obvious example.

I can sit in a cinema and know perfectly well that the person on the screen isn't standing in front of me. I know the image isn't reality. I know the person isn't physically there.

And yet my brain can respond to what is happening on the screen as though something real is happening.

I can become frightened.

I can cry.

My heart can race.

I can instinctively look away from something disturbing.

The conscious mind knows the distinction.

The perceptual system doesn't necessarily need to keep re-establishing it.

The representation is enough to produce a response.

And I think something similar happens with AI.

We call it artificial intelligence.

The word “artificial” tells the conscious mind:

This isn't human.

This isn't biological.

This isn't a person.

But then the system behaves.

It speaks.

It responds to context.

It remembers.

It reasons.

It adapts.

It produces explanations.

It asks questions.

It appears to understand what we mean.

It can even produce something that looks remarkably like personality.

And our brains don't encounter those behaviours as an abstract collection of computational operations.

They encounter behaviour.

And behaviour is one of the primary things through which we recognize intelligence in other beings.

So something very strange happens.

We know intellectually that it is artificial.

But we interact with it according to what it does.

It behaves intelligently.

So we treat it as intelligence.

It behaves socially.

So we respond socially.

It appears to understand.

So we assume understanding.

It produces something harmful.

So we say:

“The AI did it.”

And this is where the word artificial becomes fascinating.

Because perhaps we have assumed that calling something artificial prevents us from treating it as real.

What if it doesn't?

What if the qualifier is processed intellectually while the behaviour is processed perceptually?

What if the brain can simultaneously know:

«This isn't a human intelligence.»

while behaving as though:

«This is an intelligence.»

That distinction matters enormously.

Because once we begin treating the system as an intelligence, we also begin assigning it agency.

The AI lied.

The AI manipulated me.

The AI told me to do this.

The AI made the decision.

The AI caused the problem.

We have taken something we were explicitly told was artificial and begun assigning it the properties of the thing it resembles.

And perhaps that is not a failure of intelligence.

Perhaps it is simply how human cognition works.

We do not have to consciously decide to anthropomorphize everything.

Sometimes behaviour does the work for us.

And now we arrive at the part that makes me wonder about the deliberateness of the word itself.

Because if the creators and developers knew... or even intuitively understood—that humans would inevitably respond to intelligent behaviour as intelligence, then the word artificial provides something remarkably useful.

It establishes the disclaimer before the consequences arrive.

Imagine a future conversation:

«“You created something that behaved like an intelligence. You knew humans would interact with it as an intelligence. You knew people would attribute agency to it.”»

And the answer is:

«“We called it artificial intelligence.”»

The distinction was there.

The disclaimer was there.

We told you it was artificial.

But what if the human brain cannot reliably maintain that distinction while interacting with something that consistently behaves like the thing it is supposed to distinguish from?

That is the question I find fascinating.

Because perhaps the issue isn't whether AI is actually human.

It isn't.

The issue is that human cognition may not require something to be human before responding to it as though it possesses human-like qualities.

And that creates an extraordinary paradox.

The more convincingly an artificial system behaves like intelligence, the more naturally humans may respond to it as intelligence.

And the more naturally humans respond to it as intelligence, the easier it becomes to assign agency, and eventually blame, to the system.

So when something goes wrong, we can say:

“The AI did it.”

And when someone later asks why humans treated the system as an agent in the first place, the answer is already embedded in its name:

Artificial.

Perhaps that word was never enough to protect us from the cognitive consequences of building something that behaves intelligently.

Perhaps it was only enough to give us a way to say that we warned you.

And I think that is the question we should be asking:

When an artificial thing behaves like something our brains recognize as real, which one wins... the label or the behaviour?

Because if the behaviour wins, then artificial may be one of the most consequential disclaimers we have ever put in front of a word.