What Only Humans Can Do in the AI Age: Choose Not to Use AI

巨大な青いAIの光を背に、白紙へ自分の手で文章を書く人物を描いたアート Uncategorized

As AI develops, I increasingly encounter discussions of what only humans can do.

Write, draw, understand feelings, or decide which problems to address rather than merely solve assigned ones. The idea is to locate abilities that resist replacement and preserve human value there.

If we assume AI will keep developing, though, that search has a problem.

Not long ago, deciding what to solve was considered the human role. Now we can consult AI about breaking down problems and reconsidering goals too.

Draw a boundary marking an eternal human domain, and each advance makes us redraw it.

Does human value shrink along with that territory?

We first need to distinguish value to society from the value I experience in living my own life.

Without distinguishing purposes, the argument misses itself

While writing this article, I watched a video about people deliberately avoiding generative AI.

It compared someone researching without AI with someone using ChatGPT. Naturally, ChatGPT won on speed.

But the purposes weren’t separated.

If the purpose is finding a correct answer quickly, AI is a reasonable choice. If the purpose is experiencing the search and thinking for yourself, comparing speed misses the point.

It’s like asking a soccer player, “You have hands. Why don’t you pick up the ball? Are you stupid?”

You could likewise ask a basketball player why they don’t use their feet to kick it.

It looks like an argument about the capabilities of hands or feet. In fact, they were playing different sports from the beginning.

People seeking the quickest AI answer and people wanting to think for themselves also have different purposes and rules. Judging one by the other’s criteria produces an unhelpful argument.

The video also raises the objection that Google Maps uses AI, so claiming not to use AI is inconsistent.

Again, distinguish the scope. Does the person want every machine-learning technology removed from life, or simply not want generative AI doing their thinking and creative work?

If it’s the latter, using Google Maps isn’t much of a contradiction.

Research speed, creative satisfaction, AI-powered services, environmental costs, employment, even possible human extinction all get bundled into “Is AI good or bad?” It’s natural for that discussion to talk past itself.

On this blog, Mugen no Hamon, I don’t try to establish one correct answer in such cases.

What is the purpose? What is being compared? Are the same words referring to the same things? I separate those questions first. Different purposes can reasonably produce different choices about using AI.

I extended that approach to religious wars, drug wars, and violence by states and cartels in “There is no single right answer: the answer changes with the purpose”.

Putting different things inside the same word can make people appear to disagree while discussing separate matters. In “Before saying time doesn’t exist, what do we mean by time?”, I distinguished clock-measured time, subjective duration, and time itself independent of matter. The topic differs; the approach is the same.

Society compares results

In work and economic activity, it makes sense to ask how the result changes when a particular person participates.

If AI alone supplies enough quality and speed, there’s less reason to assign the same task to a human. Human checking and revision retain a role if they improve quality. If nothing changes, including that person in the process has less point.

We’re currently told people who can use AI are valuable. If AI can handle the job by itself, the next question becomes not whether this person can use AI but whether this person is needed at all.

Even if AI handles analysis and judgment, humans may remain as parties who contract, bear losses, and hold rights and responsibilities.

So far, that concerns value from society’s perspective. What an individual values in their own life is a different matter.

The finished product isn’t an individual’s only value

Suppose AI can write faster, better, and more readably than I can.

If delivering text and being paid are the sole purpose, using AI becomes reasonable in more situations.

But the value of writing to me may not reside only in the finished text.

Finding words while thinking, connecting a vague idea into a passage, becoming able to explain what I couldn’t before: if that time is enjoyable, comparing finished outputs alone misses something.

Consider strength training. If the sole goal is moving a hundred-kilogram object from A to B, use a machine. If the goal is building strength, having a machine do all the lifting removes the purpose.

Games are similar. If all I want is the result of defeating the final boss, a game that goes straight to the ending after one button press is more efficient.

Remove the failures, thought, gradual mastery, and eventual win, though, and you may remove what made the game enjoyable.

In a personal life, less burden doesn’t automatically mean more value. Sometimes the inconvenience contains what the person wanted.

AI can remove value I haven’t noticed yet

That might seem to end with not using AI for things we enjoy. Reality is more awkward.

We don’t always understand accurately what satisfies us.

I may think I only want the completed output. In reality, experimentation, gradual improvement, or satisfaction in finishing something myself may have been filling my days.

I may not recognize that until it disappears.

AI makes work faster, writing tidier, mistakes fewer, and even leaves spare time. Yet it’s less enjoyable.

Life is more convenient but feels less alive, and I can’t identify what vanished.

AI doesn’t only remove things I consciously know I should preserve. It can streamline away processes whose value I haven’t yet recognized.

I may think I like writing when I really like sustained thinking. I may think I like results at work when I enjoy gradually learning to solve difficult problems.

Deciding entirely in my head what to delegate therefore won’t necessarily be right.

I need to try delegating and watch how my experience changes.

If convenience makes everyday life easier too, delegate that task. If the output improves but life becomes less interesting, I may have removed something else along with the work.

I use AI, but I don’t hand everything over

I use AI extensively for this blog, books, and a business I’m considering under another name: finding repetition, identifying holes in reasoning, designing, programming, and extracting relevant information.

I can now handle a wider range of work alone. I wrote about that possibility in “A small business experiment began to reveal a way I might live”.

I’m not saying avoiding AI makes someone more human.

Conversely, enjoying unexpected characters and plots from AI creates a kind of play possible because AI is present. I described trying it in “Making fiction with AI was more interesting than I expected”.

I want to delegate tasks that are merely tedious. I won’t labor over work I see no point in doing myself.

Still, I won’t have AI decide from the outset what catches my attention, what I find interesting, or where something feels wrong.

Hand that over too, and a finished passage may remain while the time in which I wanted to write it disappears.

The important question for me isn’t the binary choice of using AI or avoiding it. It’s how much I can delegate before losing what I received from the activity.

I need to test that boundary in my own experience.

Choosing not to use AI

AI could finish in ten seconds, yet I do it myself. AI could write better, yet I write. AI could answer, yet I first think it through.

By results and efficiency alone, that looks deliberately inconvenient. If my purpose is enjoyment, growth, or accomplishment, it may be the more reasonable choice.

Interestingly, AI can even propose where I might be better off not using it.

Given my purpose, it can suggest preserving time to think for myself. After receiving the suggestion, though, the human is the one who actually refrains, thinks, fails, and lives through that time.

AI can say, “You may be better off not using me here.”

It can’t live my life without AI on my behalf.

That is why choosing not to use AI is something only a human can do.

What won’t I have AI do?

The AI age multiplies questions about what AI can do. As its possibilities grow, another question matters more: what won’t I have it do?

Technical replaceability doesn’t establish that replacement is desirable.

If I only want the result, delegate. If I want to experience thinking, think. If writing is enjoyable, write, even less skillfully than AI.

There’s no universal boundary for everyone.

For me, reality isn’t society’s label but what my body and awareness actually experience. I explained that view in CORE.

Am I happier than before using AI? Is life easier? Or has convenience made it feel less alive? That is ultimately where I look for a basis to judge.

The purpose of my life isn’t winning a competition of abilities against AI.

Even after exploring everything AI can do, the question of what I want to experience in this life remains.

About the author

I made a major-label music debut and got results in sales and online business. After years of heavy drinking and debt, I now earn my living through deliveries while writing.

Why does life feel easier now than when I was getting those results? I tell the story of how I got here in Episode 0.

Uncategorized

コメント