AI watermarking doesn't worry me

This week I read about the new watermark Anthropic is adding to text generated by Claude. And a fairly predictable worry showed up: "How do I hide that I used AI now?".

I see it the other way around.

The problem was never the AI

The difference is in who directs it and with what judgment they use it.

An AI can produce something mediocre or become an extraordinary tool. It depends on what's behind it: the idea, the instructions, the decisions, the corrections, the knowledge and the experience of the person using it.

A technical detail I find especially interesting

According to Anthropic, detecting this watermark does not prove that Claude is the author of a text. It only makes it possible to estimate that Claude took part in generating or editing it.

It doesn't say who had the idea. It doesn't say who designed the structure. It doesn't say who made the decisions or who provided the authorship.

Even a text written by you and later translated by Claude can end up marked. The watermark certifies the pass through the tool, not the authorship of the ideas.

What do I gain by using AI well?

I can research, build, review, translate, document and carry out certain tasks in a fraction of the time.

And I don't call that cheating. I call it productivity.

The judgment is still mine

Let the watermark stay. I have nothing to hide about using AI. The value of my work isn't in pretending I do every task without tools.

It's in knowing what I want to build, knowing how to direct them, spotting when they get it wrong, correcting them and deciding what works and what doesn't.

The tool can leave its mark. The judgment is still mine.