The Thinking I Don’t Want AI to Do for Me

I’m using AI more and more throughout my day.

I use it for work. I use it to answer random questions. I use it to provide feedback on my writing, suggest edits, organize my thoughts, and help me communicate more clearly.

It is incredibly useful.

But the more I use it, the more I’m left with a nagging question:

Is AI slowly making me dumber?

Or is it just making me lazier?

Or, in the best-case scenario, is it simply making me more effective?

When the Machine Does the Thinking

At work recently, I was developing a document that captured several organizational priorities. For each priority, the document included a brief description of the issue and a set of proposed solutions.

I wanted to add another section explaining the anticipated impact of those solutions.

Not long ago, I would have sat down at the keyboard and written it myself. I would have drawn on my knowledge of the organization, considered what each effort was intended to accomplish, and tried to turn those ideas into clear language.

The first draft probably would have been rough.

I would have revised it a few times and then sent it off for review.

This time, I did something different.

I fed the draft into my company’s AI tool and asked it to write the new content for me.

The result came back almost instantly.

And it was excellent.

I reviewed the language and made a few small adjustments, but honestly, it was probably 95 percent of the way there.

The problem was that I had barely needed to think.

I had produced a better result in a fraction of the time, but I had also robbed myself of the intellectual work required to create that result on my own.

I’m not sure I would have written anything better.

But that isn’t really the point.

The point is that I didn’t even try.

I let the machine do the thinking for me.

Sometimes the Thinking Is the Training

We generally accept that muscles weaken when we stop using them.

The same may be true of certain mental abilities.

Writing, reasoning, remembering, navigating, and solving problems are not just ways of producing answers. They are forms of practice.

Sometimes the answer is the product.

But sometimes the thinking required to reach the answer is the training.

When we skip that process, we may save time while also giving up the exercise that helps us become more capable.

That does not mean every question deserves a long intellectual struggle.

There is no particular virtue in spending three hours finding a simple fact that AI can provide in three seconds.

But not all thinking is the same.

There is a difference between asking AI to retrieve information, asking it to sharpen ideas I have already formed, and asking it to form the ideas or judgments for me.

The first often makes us more informed.

The second can make us more effective.

The third may be where we need to be more careful.

Level 1 Thinking - Finding Information

While visiting my mom’s house in the woods of Maine, I recently took a walk down to the farm pond.

The walk started pleasantly enough, but along one stretch of the path I was suddenly surrounded by horseflies. They buzzed around my head, followed me for a short distance, and then disappeared almost as quickly as they had arrived.

I’ve taken this same walk for years, and I remembered having the same experience many times before.

The horseflies always seemed to bother me along that one particular section of the path.

It made me wonder: Are horseflies territorial?

I pulled out my phone, asked ChatGPT, and had an answer almost immediately.

The answer explained that horseflies may repeatedly patrol areas where they are likely to find animals or people.

Who knew?

In this case, AI made me slightly more knowledgeable.

But I also started thinking about how I might have answered the same question at different points in the past.

A few years ago, I probably would have searched Google. I would have opened several articles, skimmed through them, and eventually found an answer.

Before the internet, I might have gone to a library, looked through an encyclopedia, or found a book about insects.

Each earlier method would have taken longer.

But the time savings is not the only difference.

In searching for the answer, I might also have encountered information about other insects, animal behavior, habitats, or ecosystems. I might have followed an unexpected reference or stumbled across a completely different subject.

AI gave me exactly what I asked for.

Nothing more.

No question went unanswered, but no unexpected knowledge was uncovered either.

Curiosity used to have more friction.

That friction could be inconvenient, but it also created opportunities for wandering, discovery, and surprise.

AI removes much of that friction.

At this level, AI is clearly making me more efficient and often more knowledgeable.

But efficiency can narrow curiosity.

The faster I receive the exact answer, the less likely I am to discover something I did not know enough to ask about.

Level 2 Thinking - Sharpening My Thinking

Writing is one area where I feel the tension most clearly.

My first drafts are usually rough. My thoughts come out slightly garbled, the structure is uneven, and the point I am trying to make may not yet be obvious.

That is part of the process. The first draft gets the ideas out of my head. The harder work often comes afterward.

Editing requires me to decide what I actually mean, remove what does not belong, strengthen weak arguments, improve transitions, and rewrite sentences that sounded much better in my head.

Editing your own work is also extremely difficult.

Most of us cannot afford to hire a personal editor every time we write an article, report, or important email.

This is where I have found AI to be extraordinarily useful.

It can identify unclear passages, suggest cleaner language, challenge weak assumptions, point out repetition, and help me see structural problems I missed.

It can even offer a little encouragement along the way.

In that role, AI does not necessarily replace my thinking.

It can sharpen it.

I still begin with my own experience, ideas, and rough draft. The AI acts more like an editor or sparring partner than an author.

Could I develop more editing skill by doing all of that work myself?

Probably.

But I also know that I could not edit my work as effectively as the AI can, at least not without far more time and effort.

The ideas are mine. The tool helps me test, refine, and express them. That feels different from asking it to generate the thinking in the first place.

Level 3 Thinking - Developing My Own Ideas and Judgment

I am comfortable asking AI to retrieve a fact. I am also comfortable asking it to identify weaknesses in something I have written or organize information in new ways.

But I want to be much more cautious when the value of the task lies in forming the original idea or judgment.

What do I believe?

What is really causing this problem?

What should we prioritize?

What tradeoffs matter?

What kind of life do I want to build?

Those are not merely questions to be answered.

They are questions that shape the person answering them.

The document I created at work fell into this third category—or at least came uncomfortably close to it. The anticipated impact of an organizational initiative is not merely a fact to retrieve or a paragraph to polish. It requires judgment. It requires understanding the organization, thinking through cause and effect, and deciding what outcomes actually matter.

AI produced strong language, but I had allowed it to form the initial judgment as well as express it. That is the part that bothered me.

The Thinking I Want to Keep

AI is extraordinarily useful, and it is not going away. The question is no longer whether I will use it, but how intentionally I will use it.

I’m happy to let AI help me find information, organize my thoughts, and communicate more clearly.

But I do not want it deciding what I believe, which tradeoffs matter, or what ideas are worth pursuing.

Those are the parts of thinking that shape who I am.

AI can help me reach answers faster. But I do not want speed to become the only measure of value.

Sometimes the thinking is not an obstacle standing between me and the result.

Sometimes the thinking is the point.