There was a period when I thought I finally understood how this collaboration worked. I wasn’t using AI blindly. Not out of laziness, and not to have it think instead of me. My questions were properly formulated - precise, measured, carefully framed. I asked for outlines and discussion materials, not finished, polished texts. I knew what I wanted from it, and I also knew what I didn’t.

I believed this already counted as conscious use.

What I didn’t notice was that something else was happening in parallel. Not with the content, not with the arguments or the professional accuracy, but with the way the material was being born - how it looked, how it felt.

There was a piece I submitted. It wasn’t bad. On the contrary. It was solid and clean. Every sentence sat exactly where it should. The transitions didn’t stick out. There were no unnecessary detours or stumbles. Even that small human uncertainty that usually characterizes me was missing.

And that was precisely the problem.

A trained eye spots this immediately: the rhythm, the smoothness, the overly precise emphasis, the flawless structure. The absence of those few sentences you later feel embarrassed about, yet still leave in because they came from you. There was no excess anger, no over-explained justification, none of that quiet “why did I phrase it like this?” feeling.

It was too perfectly in order.

My boss didn’t argue. He didn’t edit it. He didn’t ask questions. He didn’t request a new version. He simply sent it back. He said, “This isn’t what it’s for, my friend. AI isn’t here to replace your work. If it could do that, you wouldn’t be working here anymore. And if you keep going like this, you’re putting both my head and your own at risk.” There was no anger in his voice, no condescension. Just routine. It was clear he had seen many cases like this before. This wasn’t only about my text. It was obvious he had encountered AI-generated work countless times. He recognized it instantly.

That was when I understood that the real question wasn’t whether I had used AI. It was whether it showed. That my style had begun to disappear. Not replaced outright, but blurred, dissolved into a language that was correct, yet not mine.

I didn’t notice it in time, because from the inside it felt efficient. Faster. Cleaner. With less internal friction. I didn’t have to walk those familiar detours where you first get confused, then gather information, clarify, and finally decide. Where you wrestle with the task. Where the success is yours - and if it fails, that’s yours too. That’s the natural order.

Standing there, I realized I hadn’t just outsourced my questions. I had outsourced my face as well. The part that makes a piece recognizable. The reason someone can say, “This was written by him.”

This was no longer a technical issue. It was existential, ethical, and personal. A hard boundary.

It became clear that it’s not enough to watch what AI does. You also have to watch what it does to you indirectly. Because when the human voice disappears—the stumble, the unnecessary detour—you don’t become better. You become replaceable.