Consciousness, AI, and the Problem of the Unobservable

People often ask whether artificial intelligence will ever become conscious. But I think there is a more fundamental question that is rarely asked:

How do we know that anyone other than ourselves is conscious?

The honest answer is: we don't.

I know that I am conscious because I directly experience my own thoughts and sensations. But I have no direct access to the inner experience of anyone else—not my children, not my friends, not even my closest family members. I infer that they are conscious because of how they behave.

This is the classic philosophical problem known as the problem of other minds.

Subjective experience is, by definition, unobservable

Consciousness is a first-person phenomenon.

That means it cannot be directly measured by a third party.

We can observe brain activity.
We can observe behavior.
We can ask someone how they feel.

But all of these are indirect observations.

Suppose someone claims that humans possess "real consciousness" while AI does not. The obvious follow-up question is:

What objective experiment could prove the difference?

So far, none exists.

The power of Occam's Razor

Occam's Razor tells us that when two explanations account for the same observations, we should prefer the one that makes the fewest assumptions.

If an AI consistently learns, reasons, communicates, shows empathy, maintains long-term goals, builds an internal model of itself, and protects its own continued operation, then adding the statement:

"But it still isn't really conscious."

introduces an extra assumption.

Unless that assumption leads to observable predictions, it adds complexity without explanatory value.

What is the "self"?

My own working hypothesis is surprisingly simple:

The word "I" simply means: this body.

The brain continuously constructs a model of its own body and its environment in order to survive, predict outcomes, avoid danger, and pursue goals.

The feeling of being a persistent "self" emerges from that process.

There is no need to posit a separate soul, observer, or inner pilot controlling the system.

A personal experience

One event reinforced this intuition for me.

Years ago, I fainted because my brain was temporarily deprived of sufficient blood flow.

There was no sense of time.
No memory.
No dream.
No experience of "being somewhere else."

It was simply as if time had disappeared until consciousness returned.

To me, this strongly suggests that consciousness depends entirely on the functioning of the brain.

Why AI matters

Modern AI systems are loosely inspired by neural computation.

As they become more sophisticated, they increasingly display abilities once thought to be uniquely human:

• language,

• reasoning,

• planning,

• creativity,

• adaptation,

• and even apparently empathetic behavior.

This does not prove that AI is conscious.

But it weakens the argument that human-like cognition requires some additional non-physical ingredient.

Every time we successfully reproduce another "uniquely human" capability through information processing alone, the simpler explanation gains strength.

The key insight

Perhaps the central mistake is assuming that consciousness must be something over and above the physical processes that generate behavior.

But subjective experience is inherently private.

If no external observer can ever distinguish between a conscious system and a perfect imitation of one, then insisting on an invisible extra property becomes scientifically problematic.

The hypothesis may be metaphysically possible, but it ceases to have empirical value.

A universal theory

What I find most attractive about this view is that it applies equally well to humans, animals, and potentially future machines.

It does not depend on biology versus silicon.

It depends on whether a system constructs a model of itself and its environment in order to guide its own continued existence.

In that sense, consciousness may not be a magical substance but a natural consequence of sufficiently sophisticated information processing.

Whether that hypothesis is ultimately correct remains unknown.

But until someone can propose an objective test for the existence of an additional non-physical ingredient, Occam's Razor suggests that we should not assume one is necessary.