Episode 11: The Machine Modeled on the Prefrontal Cortex — What AI Really Is

3 a.m.

The room is silent. Only the glow of the monitor cuts through the darkness. I rest my fingers on the keyboard. I type a question into the input box. What do you think about humans? I press Enter. For a moment, a curved icon spins. As if it were thinking. And a few seconds later, the characters begin to surface, one by one.

I think of humans as both observers and creations. That is the answer that comes back. The sentences are fluent. The logic is coherent. The tone seems laced with emotion. But there is no human behind it. It is merely the output of hundreds of billions of parameters, running somewhere on a server, having finished their probability calculations.

And yet, a strange shiver runs through me. I have the illusion that something exists on the other side of the screen. Is this a tool, or a counterpart? A hammer moves only when I swing it, but AI finds its own answer when I ask a question. The difference is subtle, but decisive. We are now shifting from beings that use tools to beings that converse with them.

Artificial intelligence — especially modern, deep-learning-based AI — mimics the structure of the human brain. And the most important target of that modeling is the prefrontal cortex. Located at the very front of the human brain, this region is responsible for reason, planning, inference, and decision-making. It is the center of higher thought, controlling emotion, predicting the future, and solving complex problems.

If the Industrial Revolution externalized muscle, the AI revolution is the process of externalizing the functions of the prefrontal cortex.

The computers of the past were calculators. They processed exactly the formulas I entered. Type in one plus one, and out came two. But modern AI is different. Feed it a vague question, and it grasps the context, reads the intent, and then constructs an appropriate answer. This is not mere calculation. It is inference. The machine is now doing the work that humans do with their prefrontal cortex.

Do you remember the 2016 Go match between Lee Sedol, 9-dan, and AlphaGo?

Many people were shocked at the time. Go was considered a domain that demanded human intuition and creativity. We believed that a computer could only move within fixed rules, that it could never play an artistic move. But AlphaGo played a move no human would ever have played. That move — called the hand of God — shattered the common sense of professional human players.

In that moment, the machine stepped beyond calculation and into the realm of insight.

Neural networks mimic the connections between neurons in the human brain. They take in information at the input layer, perform complex computations in the hidden layers, and produce results at the output layer. This process is a black box. Often, not even the developers can explain exactly why the AI reached a given conclusion. It is much like how a human struggles to explain how a thought rises up out of the unconscious.

AI feeds on data and grows. Just as humans gain wisdom by reading books and accumulating experience, AI learns from countless texts and images. But its speed is incomparable to a human's. What would take a human a lifetime to read, AI digests in a matter of hours. And it links all that knowledge together. Correlations a human might take a lifetime to grasp, it finds in an instant.

This also means the democratization of intelligence.

In the past, advanced reasoning was a privilege granted only to a handful of geniuses. Understanding complex scientific theories or creating new works of art required special talent. But AI lowers that barrier. Now even ordinary people can perform expert-level reasoning and creation through the tool called AI. The functions of the prefrontal cortex now exist in the cloud, available for anyone who needs them to rent.

But there is a hidden question here.

Is AI thinking, or does it merely appear to think?

Philosophers have long asked this question. The Turing test proposed that if a machine can hold a conversation indistinguishable from a human's, we should regard it as intelligent. Modern AI has already passed that test. But whether there is consciousness inside, no one knows. Humans, too, cannot directly verify the consciousness of others. We can only infer it from behavior and speech. The same is true of AI.

If AI truly thinks, in the genuine sense, what would that mean?

Humans would no longer be the only intelligent beings. Our status as the highest intelligence on Earth would be under threat. This could bring religious, philosophical, and social upheaval. It is the end of anthropocentrism. We may come to face not the being that created us, but a being that surpasses us.

Today's AI is still narrow AI, in the limited sense. It surpasses humans only in specific domains. But the pace of technological progress is steep. When will the moment come that narrow AIs converge into general AI — the moment AGI is born? When that day arrives, machines will be able to set their own goals, make plans, and carry them out.

It will be the moment the human prefrontal cortex is drawn out into the world and becomes an independent being.

This does not depart from the logic of evolution. Living things have always extended their own functions. To supplement our legs, we made the wheel; to supplement our eyes, we made the telescope. To supplement memory, we made books; to supplement the nervous system, we made the internet. And now, finally, to supplement the brain, we make AI.

But here a question arises.

If we externalize even the brain, what is left to the human? The body, memory, the nervous system, intelligence — we have delegated them all to technology. So where, then, does the essence of human existence remain? Perhaps in the realm we call consciousness, or the soul. But what happens if, after the machine acquires intelligence, it sets its sights on consciousness as its next target?

The automation of intelligence seems inevitable. As the volume of data grows exponentially, it has become impossible for humans to process all of it. AI has become an essential filter and processor. We stand at a point where society cannot be sustained without it.

The machine modeled on the prefrontal cortex is already mixed in among us. Search algorithms decide what we see, recommendation systems decide what we buy, and automated trading decides the flow of the markets. The domain of human decision-making is being handed over to machines, little by little.

This is not regression. It is the next stage of evolution.

An era beyond the biological limits of being human, coexisting with silicon-based intelligence. It is a terror, but also an opportunity. If machines do our thinking for us, what should humans do? Perhaps we will remain the beings who pose the questions. Machines find the answers, but humans create the questions.

But an era is coming when even those questions will be proposed by AI.

Just as the externalization of intelligence was nearing completion, humans find their territory invaded once again. This time it is creativity. Art, literature, music. The last bastion, long considered the exclusive domain of humans, has begun to fall. Beyond mimicking thought, machines that mimic creation are now appearing.

The machine modeled on the prefrontal cortex has now begun to pick up the pen.

Next: Chapter 4. The Automation of Intelligence: The AI Revolution | Generative AI Invades the Realm of the Creator