For months, I treated AI like a shortcut. Then I realized I was losing something I didn't expect.
A few weeks ago, I found myself opening ChatGPT to answer a question I probably could have answered on my own.
It wasn't an important question.
I wasn't researching a complex topic or solving a difficult problem.
I was simply trying to come up with a title for an article.
Before I had written down a single idea, I opened ChatGPT.
Ten seconds later, I had a list of suggestions.
The problem wasn't that the suggestions were bad.
The problem was that I never gave myself a chance to think.
That small moment stayed with me longer than it should have.
And the more I paid attention, the more I noticed the same pattern repeating throughout my day.
Need an email?
ChatGPT.
Need an outline?
ChatGPT.
Need ideas?
ChatGPT.
Need a summary?
ChatGPT.
Need a different way to phrase something?
ChatGPT.
At some point, I had stopped using AI as a tool and started using it as my first thought.
That's when I decided to try a small experiment.
For one month, I would stop asking ChatGPT everything.
Not completely.
I would still use it when it genuinely helped.
But before opening it, I had to spend time thinking for myself first.
I didn't expect the experiment to change much.
I was wrong.
When Convenience Becomes a Habit
The most powerful technologies don't change our lives overnight.
They change our habits quietly.
Think about GPS.
Most of us can still read a map.
We just don't.
The same thing happens with calculators.
We can do basic math in our heads.
We simply choose not to because something faster exists.
AI feels similar.
The first time you use it, it feels like a productivity boost.
The hundredth time you use it, it starts becoming a default behavior.
That's not necessarily bad.
The danger is that habits become invisible.
You stop noticing them.
I didn't notice how often I was outsourcing small moments of thinking until I deliberately stopped.
And once I noticed, I couldn't unsee it.
The Lost Value of Sitting With a Problem
One of the first things I rediscovered was something surprisingly uncomfortable.
Thinking is slow.
When you don't immediately reach for AI, there is often a period of uncertainty.
You stare at a blank page.
You struggle with an idea.
You wonder whether you're approaching a problem the right way.
For years, I treated that feeling as a sign that something was wrong.
Now I think it might be one of the most valuable parts of the process.
Many of my best ideas have never arrived instantly.
They appeared while walking.
While taking a shower.
While staring out a window.
While wrestling with a problem long enough for my brain to connect ideas in unexpected ways.
AI can shorten that process.
Sometimes that's useful.
But sometimes the struggle is where the insight lives.
The Day I Almost Missed a Great Idea
One afternoon, I was outlining a new article.
My first instinct was to ask ChatGPT for ten angles on the topic.
Instead, I forced myself to spend twenty minutes brainstorming alone.
The first few ideas were terrible.
The next few were predictable.
I started wondering if I was wasting my time.
Then something unexpected happened.
Around fifteen minutes in, a completely different angle appeared.
It wasn't something I had seen before.
It wasn't obvious.
It connected two unrelated ideas in a way that felt interesting.
It eventually became one of my most-read articles.
Would ChatGPT have suggested it?
Maybe.
Maybe not.
But that's not the point.
The point is that I almost skipped the process that led me there.
Faster Isn't Always Better
We live in a culture that celebrates speed.
Fast answers.
Fast results.
Fast content.
Fast decisions.
AI fits perfectly into that mindset.
And to be fair, speed can be incredibly valuable.
But speed and quality aren't always the same thing.
Sometimes the fastest answer is also the most obvious answer.
The most common answer.
The least original answer.
Original thinking often takes longer.
Not because it's complicated.
Because it requires space.
Space to question assumptions.
Space to explore alternatives.
Space to be wrong for a while.
That's harder to do when an answer is always one prompt away.
What Changed When I Used AI Less
The interesting part wasn't that I became more productive.
In some ways, I became slower.
At least initially.
But several unexpected things happened.
I started writing more original ideas.
I became better at asking questions.
I trusted my own judgment more.
I spent less time looking for the "perfect" answer.
And I enjoyed the creative process again.
That last one surprised me most.
I hadn't realized how much of my work had become optimization.
Every task felt like something to complete as efficiently as possible.
The experiment reminded me that creativity isn't always efficient.
Sometimes it's messy.
Sometimes it's frustrating.
And sometimes that's exactly why it's valuable.
I Still Use ChatGPT Every Day
This isn't an article about quitting AI.
I still use ChatGPT almost every day.
I use it for research.
I use it for brainstorming.
I use it to organize information.
I use it to improve drafts.
It's one of the most useful tools I've ever used.
What changed is the order.
Before, AI was my first step.
Now it's usually my second.
I try to think first.
Then I use AI.
That small change has made a bigger difference than I expected.
Because AI works best when it helps develop your ideas—not replace them.
Three Rules I Follow Now
The experiment led me to three simple rules.
Rule 1: Think Before You Prompt
Spend a few minutes with the problem first.
Write down your own ideas.
Then ask AI.
You'll often get better results because you'll ask better questions.
Rule 2: Use AI to Expand, Not Replace
Instead of asking for answers, ask for perspectives.
Ask for alternatives.
Ask for weaknesses in your thinking.
Use AI as a thinking partner rather than a thinking substitute.
Rule 3: Protect Time for Independent Thought
Not every problem needs an instant solution.
Sometimes the best thing you can do is sit with a question a little longer.
The answer may arrive when you least expect it.
The Real Lesson
I started this experiment because I was curious about how much I relied on AI.
I finished it realizing the issue wasn't AI at all.
The issue was how easily convenience can become dependency.
Technology is most useful when it expands our abilities.
But every tool also changes our habits.
The challenge isn't deciding whether to use AI.
The challenge is deciding when not to.
And that might become one of the most important skills of the next decade.
The question I keep coming back to is this:
If AI can answer almost anything, how do we make sure we don't stop asking our own questions first?