Corentin lives in New York.

We mostly talk through voice notes.

A few days ago, he sent me one after reading my latest essay.

“I’ve never been officially diagnosed,” he said, “but I’m almost certain I’m dyslexic.”

Then he laughed.

“For years, writing a cover letter could take me weeks. Not because I didn’t know what I wanted to say.

Because I could never find the words that matched what I was actually thinking.”

Today, he uses artificial intelligence.

Not to write for him.

To search.

For a phrase.

A nuance.

The right words.

Toward the end of his message, he told me about something that had happened earlier that day.

He had been revising a letter for more than an hour.

ChatGPT suggested a sentence.

He deleted it.

Another one.

No.

Another.

Still no.

Then he stopped.

“That one…”

Silence.

“That’s exactly what I meant.”

It probably wasn’t the best sentence.

But it sounded like him.

As I finished listening to his voice note, I realized I do exactly the same thing.

I started wondering how many sentences, throughout my life, I had kept simply because they sounded good.

How many times had I chosen an elegant sentence over a faithful one?

I’m not just talking about books.

I thought about cover letters.

Emails.

Conversations.

All the times I’d tried to say what I thought people wanted to hear instead of what truly reflected my own thinking.

Over the past few months, I’ve noticed a new way people are using artificial intelligence.

Some ask it how to break bad news.

Others describe an argument with their partner.

Some search for the right words before a difficult conversation.

Others simply try to understand why a single comment hurt them so deeply.

For some people, that’s unsettling.

Sometimes even worrying.

“It’s not human.”

“It’ll never replace a therapist.”

I understand those reactions.

But I also feel they miss something important.

I’ve never met a good therapist who could live your life for you.

I’ve never met one who tells you who you are.

The best ones do something else.

They listen.

They reframe.

They ask a question you had never thought to ask yourself.

And suddenly, a story you’ve been telling yourself for years becomes a little less convincing.

Sometimes I wonder if this is how many people are beginning to use artificial intelligence.

Not to receive answers.

But to keep searching.

I often think back to Corentin.

To that hour he spent on a single sentence.

From the outside, it looked like wasted time.

He knew he was looking for something else.

Not a more elegant sentence.

A more faithful one.

People often say that artificial intelligence will eventually make us all sound the same.

I’m no longer convinced that’s the real risk.

The real risk may not be talking to a machine.

It may be accepting the first answer too quickly.

The first idea.

The first story.

The first version of ourselves that happens to work.

I think about Corentin again.

For more than an hour, he wasn’t searching for a better sentence.

He was searching for one that sounded like him.

I think many of us spend a good part of our lives doing exactly the opposite.

Then, every once in a while, something happens.

We reread a sentence.

We hear our own voice.

And almost with surprise, we whisper to ourselves:

“That’s what I meant.”

— Kevin