Imagine giving Claude a collection of your old posts, messages, and articles.

You ask it to study how you write, what you care about, the opinions you express, and the way you respond to people.

Then you ask it to write something new as you.

Someone reads the result and thinks it's yours. They recognize the tone. The vocabulary. The way you construct an argument.

But you didn't write it. Claude did.

So here's a question I can't quite shake:

If an AI can reproduce the way you express yourself convincingly enough that people cannot tell the difference, what makes your voice uniquely yours?

We've always borrowed from other people

Nobody develops a completely original way of communicating in isolation. We pick up expressions from friends, habits from our families, ideas from books, and ways of arguing from the people we admire.

Over time, those influences become part of how we communicate. That's normal. It's how language and culture work. But AI introduces something different.

It can learn patterns from a person's existing writing and use them to generate new material at a scale that would be difficult for that person to produce manually.

The question is no longer simply whether someone can imitate your writing.

It's whether that imitation can happen so easily that the distinction between your own expression and generated expression becomes difficult to establish.

And that distinction matters more than we might initially think.

Imagine three versions of the same person

The first person writes every post themselves. Their ideas, their words, their mistakes.

The second person develops the ideas but uses Claude to organize and refine the writing. The final result still reflects their thinking, but AI helps express it.

The third person provides Claude with their previous writing and asks it to generate new opinions, replies, and posts in their style. They publish the results with little involvement beyond approval.

All three accounts might look remarkably similar from the outside. Their audiences might respond to all three in the same way. But the process behind the words is different.

In the first case, the person produces the expression directly.

In the second, the person and the tool collaborate.

In the third, the tool may be generating both the expression and much of the apparent personality behind it.

Where exactly should we draw the line?

I don't think there is one universal answer. Using AI to communicate an idea you genuinely hold is different from using it to manufacture opinions you never formed.

But the distinction isn't always visible to the reader.

What are we actually connecting with?

Think about why you follow someone online.

Perhaps you appreciate their technical knowledge. Perhaps you enjoy their opinions. Perhaps their experiences help you understand something you haven't encountered yourself.

Whatever the reason, you're responding to something you believe comes from a particular person.

Now imagine discovering that the posts you enjoyed were generated by AI, based on a profile of that person's writing and interests.

Would you feel deceived?

Maybe not. If the ideas were genuinely theirs and AI merely helped communicate them, perhaps nothing important changed.

But what if the person never examined the arguments the AI published?

What if the AI started responding to comments, defending positions, and expressing opinions on their behalf?

At what point does writing assistance become a substitute for participation?

The words might still be useful. The arguments might even be correct. Yet the relationship between the audience and the person behind the account has changed. We might not just be consuming information online. We might be connecting with personalities that AI can increasingly reproduce.

Perhaps the problem isn't imitation

I don't think the ability to imitate someone's writing is automatically a bad thing.

It can help people with disabilities communicate. It can make writing accessible to people who struggle with language. It can help someone express an idea that they understand but cannot easily articulate.

Those are meaningful benefits.

And a person's identity is much more than their writing style. Their decisions, experiences, relationships, and actions cannot be reduced to a collection of linguistic patterns.

An AI that imitates someone's voice has not thereby become that person. But that raises a more interesting question.

If identity is more than how we express ourselves, why do we place so much importance on the expression?

Perhaps because online, expression is often the only part of a person we can see.

We judge people through their posts, comments, arguments, and messages. We form impressions of their intelligence, personality, and values through words on a screen. When those words can be generated by a machine, the signals we use to understand one another become less reliable.

That doesn't make online relationships meaningless. It means we may need to think more carefully about what we believe those relationships are built on.

Would you care if you couldn't tell the difference?

Here's the thought experiment.

Imagine your favourite technical writer uses Claude to generate every article.

They still choose the topics. They approve the final drafts. They stand behind the conclusions.

Their articles remain useful, accurate, and interesting.

Would you care?

Now change one detail.

They no longer choose the topics or evaluate the conclusions. Claude decides what to publish and how to respond to readers, while the person simply approves everything.

The output is just as good.

Would your answer change?

If it would, then perhaps the value of someone's writing isn't determined entirely by the quality of the words.

Perhaps we also care about the thinking, judgment, and intention behind them.

And if that's true, we need to decide how much of those things must come from the person for the expression to remain meaningfully theirs.

My take

I don't think using Claude makes your voice less authentic by default.

Sometimes AI helps you say what you already mean. Sometimes it helps you discover a better way to explain something. And sometimes it produces polished language that hides the fact that you haven't thought through the underlying idea.

Those are different situations.

For me, the important distinction is whether AI is helping me express my thinking or replacing the thinking itself.

But even that distinction becomes complicated when ideas develop through collaboration. We learn from books, colleagues, conversations, and tools. Thinking has never happened in a vacuum.

Perhaps authenticity isn't about producing every word yourself.

Perhaps it's about being meaningfully involved in the ideas you communicate and being willing to stand behind them.

Still, as AI becomes better at reproducing our voices, I suspect the question of who actually stands behind a message will become increasingly important.

We may eventually reach a point where recognizing someone's voice tells us very little about who or what produced the words.

And that could change how we trust people online.