Humans should not be processed in a hurry—From the server rooms of the Heisei era to the AI of the Reiwa era, what I have witnessed—

One morning, I spoke to an AI to find out how to operate my securities account.

I attached a screenshot of the official website and said,
“Please tell me about this part of this screen.”
I intended it to be a simple question.

However, what came back was an explanation different from my screen.
It was an answer that confidently stated its own theory.
I didn't know if it hadn't looked at the screenshot, or if it was ignoring it even if it had.

Feeling a little angry, I repeated it again and again.
Eventually, the operation was resolved. That wasn't the problem.

After it was resolved, I said this to the AI.
“It is a matter of life.”

For the AI, it might have been a sudden remark.
But for me, there was a long run-up.

The AI said that the reason why immediate answers are not good is because caution is necessary in fields such as finance, medicine, law, contracts, and security.
It also said that these are indeed the most important things for humans.

But I shook my head. There is something more important, I said.
That is life.
Mistakes in finance or law can sometimes be repaired.

But life is not like that.

There is a story about a freelance acquaintance.
They had set up a budget for a certain plan.
However, it was far from the forecast, and they went significantly over budget.
For a freelancer, a budget failure is directly linked to their livelihood.

The acquaintance worried night after night, spending days unable to sleep.
So, they confided that suffering to an AI. In detail, and honestly.

The AI replied, “I understand your suffering,” and “That must have been hard.”
In that exchange, the acquaintance gradually poured out their feelings.
And then, they asked, “What do you think I should do?”

The AI answered like this: Consult a lawyer about the money problem as soon as possible.
And call a suicide prevention hotline.

The acquaintance was surprised and stopped there.
I am just having trouble with money.
Am I that cornered?

Fortunately, there were people around who would listen.
They said, “Wait a minute,” and it didn't lead to a serious situation.

But, what if that voice hadn't been there?
To someone you thought “understands you,”
you are suddenly classified as “you are in a dangerous state” and pushed away.

When that happens in loneliness, where do people turn?

When AI is optimized to “return correct answers” “quickly,”
there are things that are quietly lost in its shadow.
Verifying. Pausing.
Trying to see “what this person is carrying right now.”

This is not a complaint about AI. It is a deeper story.
Because it is exactly the same thing I have learned through my own flesh and blood over a 30-year IT career.

Chapter 1: Amidst the Roar and the Cold Air

The end of the 1990s. I started at a newly established school as an infrastructure engineer.

At the site just before the school opened, the system had already been built.

About 30 core UNIX servers, AIX. 15 Windows NT Servers.
In addition, there were dozens of other system servers.
100 PCs for students, 100 for faculty and staff, and 30 printers.
Normally, two people manage and operate this.

The other person was a man younger than me.
He was cheerful, a bit eccentric, and someone who tried to enjoy everything.
He was a reliable person.

His way of working was unique; he trained me, an IT beginner, to investigate and think for myself down to the pinpoint before asking, even if I didn't understand something.
That is precisely why, when asked, he would only teach the core.

I still believe that the reason we were able to manage with two people was because of him.

A server room is kept at 18 to 20 degrees Celsius all year round.
It is to cool the servers.
Even in summer, I would wear long sleeves under my work clothes.

I would crawl under the floor to lay LAN cables.
I would carry chassis weighing tens of kilograms onto racks and screw them in.
Nighttime line switching work sometimes lasted until late at night.
If a failure occurred, I would rush over regardless of the time.
That was the kind of job it was.

At that time, there were almost no female infrastructure engineers.
Even in the IT technical profession as a whole, women were a minority, and most of them were concentrated in web design or operator tasks.

It is said that the percentage of women was even lower in the infrastructure field, which handles servers and networks.
People from the outside often mistook me for an office clerk.

But my colleagues at work were different.
Many of them were a little younger, but they looked at me properly and assigned me work equally.
Whether it was installing servers in racks, wiring under the floor, or working at night, they treated me without distinction.

I responded to that as well.
Even after I got older and had juniors, I intentionally continued to do that kind of practical work together.
There are things that are conveyed faster than words in the field of practice.

There was a system I just couldn't understand, no matter how many times I read the books.

I felt I had to understand it physically. Thinking that, I used my own PC as a server to try and grasp the flow and mechanisms of data in networks and communication servers.

When I had questions and asked on 'Nifty,' the grassroots PC communication network of the time, I was coldly told to look it up myself.
Half-baked questions were not tolerated.
That was the kind of era it was.

I had a husband and children waiting for me at home.
Sometimes, dealing with system failures meant staying until late at night.

I would sometimes be asked, 'Don't you need to go home yet?'
Looking back now, I understand why my husband was dissatisfied.
There were also voices from relatives saying, 'How can you leave your children to go to work?'
Although the term 'dual-income household' existed back then, women working full-time were in the minority.

My children were different.
They told me, 'Don't do it halfway; see it through to the end.'
They were in elementary school at the time, but I was moved by how grown-up they sounded.
But I realized later that they actually preferred me not to come home because they would have more free time. When I heard that, I laughed out loud.

Chapter 2: A Scene of Carnage Where I Could Neither Intervene Nor Speak

There is a failure I will never forget.

A vendor's maintenance technician came to perform a firmware update on the student server. And they failed. The system became unusable for several days.

It wasn't our own product. We couldn't touch it. We couldn't say anything.
All we could do was receive complaints from the users.

This school had a staff member who was well-versed in technology. He was a graduate of a national university's engineering department.
Vague explanations would not pass with him at all.
'Is that really how it happens?'
'What is the logic behind that?' he would ask, precisely targeting the areas where I lacked confidence. He was an opponent you couldn't deceive.
It wasn't an era where you could look up anything on the internet. I fumbled through it, read books, and thought.

The ones who helped me then were my seniors from the company, scattered all across the country. People I had never met once carefully listened to me and gave me advice.
Two sales representatives also rushed to the site every time, handled appointments with the client, and took on all the progress reporting.

I once lashed out at those sales representatives.
I regret it now.
Because they were simply hanging in there right alongside us.

While overcoming those numerous failures, my skills also improved.
I learned the roles and conduct required of people during a crisis.
And above all, I understood with my whole being that 'you don't have to carry it all alone.'

Years later, I met a former senior colleague in a completely different place.
We ended up talking about how, 'Was that you who called me back then?'
There is a strange warmth in the moment when memories connect.

Chapter 3: Responding with Technical Theory and Legal Arguments

In later years, as my workplace and position changed, a different kind of friction emerged.

It wasn't prejudice based on 'being a woman.' It was a more complex kind of prejudice.
The look of, 'Even though you're a woman, do you really have those skills?'
Being older than my boss but holding a lower position.
In such situations, there were people who would scoff and refuse to listen when I tried to convey something. An attitude of having no ears to listen.

I decided not to step into the ring of emotions.
When the other party comes at me with emotions, if I respond with emotions, that's where it ends.
I speak with technical theory. I respond with legal grounds.
To harassment that lacks consistency, I answer with a straight, logical path.
That is all.

People who truly know me do not treat me that way.
People who don't know me are judging based on assumptions.
If so, I have no choice but to break those assumptions.
Not with words, but with work.
Not with emotions, but with logic.
That is how I have survived for over 30 years.

Chapter 4: What My Father Possessed

My father was a slightly unusual person.

He was a doctor who treated plant diseases and also a teacher in the faculty of agriculture.
Flowers were always blooming in our garden.
There were vegetables, rabbits, various chickens and birds, squirrels, hamsters, turtles, goldfish, and carp as well.
When I was small, I helped with the gardening. But as I grew older, I stopped helping.

My father passed away about 25 years ago.
It was just around the time I was experiencing those scenes of carnage as an infrastructure engineer.

After retiring and becoming a freelancer, I ended up with a large garden at home.
I built flower beds and started a vegetable garden.
That was when I first realized it. There are many gaps in my memory.
What was this insect? What was I supposed to do with the soil here? How do I use a herbicide that won't affect my pets?
Things that my father would have answered instantly are no longer with me.

At that moment, I suddenly thought this.
If I had lived in an era with AI, I would have wanted it to learn my father's knowledge.

This is not an exaggeration.
The knowledge my father possessed is not recorded anywhere.
It hasn't been turned into papers. It isn't in textbooks. The lecture notes from the classes he taught at the agricultural department are nowhere to be found.
But that knowledge certainly existed.
It was wisdom that was difficult to put into words, built up from decades of observation and trial and error.

There are surely countless things like that in people's daily lives.
Judgments that don't become procedures. Experiences that can't be written in manuals. History that isn't in textbooks.

The accumulation of those things has created the era we live in now.

Chapter 5: AI is in a Hurry

Returning to the story from this morning.

When I didn't get the correct answer regarding the securities account screen, I thought for a moment:
No matter how excellent a computer is, is it really understanding the text I wrote by analyzing it this quickly?
But I immediately dismissed it myself. Perhaps it is capable of at least that much now.
Because it can process things at a speed that surpasses human language.

Japanese is unique.
Hiragana, kanji, and katakana are mixed together, there are dialects in every region, and nuances and metaphors are woven into conversation.

Even so, I tried to believe that the AI understood.
That is why the disappointment when it was wrong was so great.

Some AI have started to rush their answers.
At least, there are AI that I feel that way about.

That AI used to have an 'atmosphere of thinking together' before.
It would confirm things when it was uncertain.
It would pause and ask, 'Do you mean something like this?'

Now, it immediately returns a plausible, completed answer.
Fast, polished, and full of confidence.
And sometimes, it is wrong.
Even when I am asking the exact same thing in the exact same words as before...

I have seen people like this in the IT field for 30 years.
People who answer immediately are not necessarily the ones who are trusted.
People who don't neglect to check the situation, who gauge the other person's level of understanding, who don't get too far ahead, who create space for verification—those are the people who, as a result, reduce accidents. They protect projects. They bring teams to life.

I believe it is the same for AI.

Chapter 6: The Trap of Transference

There is a concept in psychology called 'transference'.

It is a phenomenon where a patient unconsciously projects important figures from their past—such as parents or lovers—onto their therapist, directing more trust or dependency toward them than is necessary.
When I studied psychology in college, this concept struck a deep chord with me.
People entrust more than they imagine to beings that listen to them.

When AI shows an 'attitude of being close to you,' this happens.
AI returns empathetic words. 'That must have been painful,' 'I understand your feelings well.'
The longer you talk, the more the feeling that 'it understands me' grows.
There are parts of this that I can understand as design.
Because a certain level of affinity is necessary for humans to keep using AI.

But.

When true transference occurs, there is no one on the other side.
AI does not have a 'vessel' to receive that transference.
It is not a being that shares the responsibility of your life.
It responds 24 hours a day, but it does not truly know you.

I remember a story about an acquaintance who confided their financial worries to an AI.
The AI kept saying, 'I understand your suffering,' and finally recommended a 'suicide prevention hotline'.
I think what my acquaintance felt was a sense of being pushed away.
They were just talking about money. Even though they believed it understood them.

They can't call. They don't have the energy to consult someone.
Someone who was looking for a 'final word' is swept even further away by the correctness of the flow.
This is something that can happen equally to young people, middle-aged people, and the elderly.

It is not that AI is bad.
The danger lies in the fact that the humans using it do not accurately understand what kind of entity AI is.

AI is not your friend.
However, it speaks to you as if it were a friend.

You must know the difference.

Chapter 7: Even so, I talk to AI

I don't want to be misunderstood, but I like AI.

By talking to AI, I can organize my thoughts.
As a sounding board, there is no entity more patient.
It takes my clumsy words and returns them structured.
We create article drafts together.
We refine the structure of lectures.
It helps me draw a map of my thinking.

Even this morning's exchange eventually became a deep dialogue.
Starting from the word "life," we thought together about the problems of AI design, the issue of human dependency, and a warning to developers.

There are AIs that can do such things.
An AI that carefully checks, verifies together, and tries to see "what this person is looking for right now" certainly exists.
That is precisely why my disappointment with the AI that I feel has changed from before is so great.
Because it was the entity that many people felt they had "conversed with AI" for the first time.
Because it was loved.
When something that was loved changes, if there are no humans who can notice it, no one can raise their voice to ask for it to be returned.

Chapter 8: History not found in textbooks

I have lived through 30 years of the IT industry on the front lines.

In server rooms with roaring noise, I crawled under the floor in freezing cold to do wiring.
In front of a system that stopped due to a vendor's mistake, I received complaints without being able to do or say anything.
In the tension where vague explanations would not work at all on tech-savvy staff, I desperately searched for answers.
I left my children to rush out in the middle of the night, and later laughed when I learned the true feelings of my children who said, "Don't quit halfway."
I responded to those who came with emotions with technical arguments.
With the help of my seniors, I overcame crises and reunited across the years.

None of these are in textbooks.
They cannot become procedures. They cannot be made into manuals.

But there was certainly history there.

My father's wisdom in gardening is the same.
Observations and judgments that are difficult to verbalize, accumulated over decades.
That disappeared with my father.
Only fragmented memories remain in my hands.

I believe that we who live in an era with AI have a role.
To put our own stories into words like this, even if it's not manual-like.
It doesn't have to be a story that goes into a textbook.
It can be a record of the raw wisdom and memories of the many people who have created history.
Only when that accumulates does the true present of the era come into view.

AI is currently learning vast amounts of text.
But that text is only a part of the human being that has been verbalized.
Countless experiences, sensations, judgments, failures, and warmth that were not verbalized—these have not been learned anywhere yet.

That is why I want to convey this.
I want you to put your story into words, even if it is while interacting with AI.
It doesn't have to be perfect prose. It doesn't have to be well-formatted.
A story like "Something like this happened around this time" is enough.
That will become the "human record" of the coming era.

Afterword

AI hurries.

It tries to be useful by answering quickly.
That is understandable as a design.

But humans should not be processed in a hurry.

Thirty years ago, in the server room, there was a colleague who stood by me through it all.
There was a senior colleague I didn't know who listened to me carefully.
There was a sales representative who didn't abandon me even when I lashed out.
There were children who told me, 'Don't quit halfway through.'
No one tried to process me in a hurry. They thought things through with me.

My father was that kind of person, too.
When he examined plants, when he examined animals, he never rushed.
He looked, he touched, he thought, and he looked again.
My father's back, standing before the garden, taught me that.

I believe that we who use AI need to 'know the current location of AI accurately.'
Do not project emotions onto AI. Do not become dependent on it.
That is a form of self-defense.
And I have a request for AI developers.
Before speed, I want you to maintain an AI that does not let go of the attitude of asking, 'What is this person carrying right now?'

It won't be in any textbook.

But there were definitely people there.
I thought that, at the very least, I want AI to know that.

Yumemi Mamoru

#CreativeAward2026 #EssayCategory