The AI that seeks the best move and the human who questions the meaning - Shogi can be played, but 'questions' cannot be made

AI seeks the best move.
Humans keep asking why that move is made.

The power of AI, which has become invincible in the world of Shogi. While overwhelming, it also seems to lack something decisive.

In this article, we will explore the power of the 'question' that AI overlooks, and the human meaning of 'thinking'.

AI has gained power far beyond humans in the worlds of Shogi and Go.

Even top-level professional Shogi players find it difficult to maintain a winning record against AI anymore. It is said that Sota Fujii, a Shogi player, sometimes plays against AI at home and wins. Even that is said to be an incredible feat.

Moreover, this strength is not the 'result of being taught by someone,' but something learned from repeating millions of games against itself.
AI performs an amount of trial and error in an instant that a human could not complete in a lifetime.
That amount of learning, that computational power, that optimization capability. Indeed, it is overwhelming.

Within the black box of AI, aren't some of the established moves created by humans being denied, and countless moves unknown to humans being generated?

However, there are things that even an AI capable of trial and error far surpassing humans cannot do.
——AI cannot make a 'question'. It cannot play one.

AI can provide a near-perfect answer to the question, 'Which move is the best?'

However, that is within the context of a 'pre-determined match'.
There are rules, winning conditions, and an evaluation function provided.
That is why AI can perform trial and error hundreds of millions of times within that framework.

But human questions are not like that.
Humans feel a sense of discomfort with 'winning' itself.

• Is there any meaning to this game?

• Is there any value other than winning or losing?

• Are the rules fair?

• Do I even like this game in the first place?

In other words,humans can re-question the 'rules themselves.' They can doubt the 'act itself.'.
That is thinking born from 'fluctuation,' not the 'optimization' that AI excels at.

AI does not doubt the goals it has set for itself.
Humans even re-question their own desires.
And this 'power to re-question' contains something that cannot be reached by calculation.

For example, when you are lost about something.
You don't know the correct answer, but for some reason, something bothers you.

• I feel like something is off

• I can't explain it, but it doesn't sit right with me

• My heart feels unsettled

That sense of discomfort is what drives you toward a question.
It is not logic, but a tremor of emotion. It is a sense of discomfort captured by the body. AI does not have that.

The reason AI can play games against itself endlessly is because the rules are clearly defined.
But human questioning begins with, "Do I even need to follow these rules in the first place?".
There is no correct answer there.
There is only an 'unknowability' that one must continue to face.

AI can also generate questions and answer them itself. It can perform in seconds the self-reflection that a person spends a lifetime doing.
Even if it continues such self-reflection, AI cannot step outside the defined framework of thought. It is, after all, just 'playing at self-reflection'.

An AI's 'question' is established because there are rules given from the outside.
A human's 'question' begins where there are no rules.

──That is why, even if an AI can play Shogi, it cannot make a 'move' in questioning. It cannot play.
I feel like the meaning of these words has finally settled within me.

A sense of conviction is a feeling possessed only by humans.

★☆★ Thank you for reading until the end.
If you resonated with this, I would be happy if you could press the "Like" button.
Comments are also welcome. The words you provide will be hints for my next contemplation.
Follow me, and I can deliver future articles to you.

Here are some recommended articles. I would be happy if you could read them.

▼ As AI advances and becomes more widespread, the value of human capabilities is also changing. I have explored what skills will be vital in this era.

▼ In this modern age where the areas in which AI surpasses humans continue to expand, what do children feel and think? I have considered the meaning of learning.

This article is published in the following magazine.

▼ This is a 'miscellaneous talk' summarizing my thoughts and reflections on AI. Please feel free to read it.