The Mirror Check
One question that keeps the human in the lead
Most AI work sessions end the same way.
You close the tab, move to the next task, and carry forward whatever the exchange produced. A draft, a plan, a decision, a direction. The work is done. You move on.
But there is a question worth asking before you close — one that takes ten seconds and changes the quality of everything you carry forward.
Did I lead this session, or did I follow it?
It sounds simple. It isn't always easy to answer honestly.
Leading means you arrived with a position. You knew what you wanted, what mattered, what was not negotiable. The AI worked against a frame you had already built. It compressed, expanded, challenged, drafted — but inside a territory you defined. The output is yours, assisted.
Following means the machine set the frame and you moved inside it. You arrived with a vague input and accepted the first structure the system offered. The categories it chose became your categories. The tone it selected became your tone. The priorities it surfaced became your priorities. The work may be good. But it belongs, in some meaningful sense, to the model.
Neither is always wrong. For low-stakes tasks — formatting, summarizing, converting — following is fine. Efficient, even. But for anything that carries your name, your judgment, your strategy, or your relationship with another person, the distinction matters enormously.
The Mirror Check is not a judgment. It is a diagnostic.
If you led: close the session. The work is yours.
If you followed: go back one step. Find the moment where the machine took the wheel. Restate your actual position — even briefly, even imperfectly — and make one deliberate decision before you move forward. Not a re-do of the whole session. Just the retrieval of your authorship.
That retrieval is the practice.
Over time, the check becomes faster. You start to feel the difference in real time — the slight drift that happens when you stop directing and start reacting. You catch it earlier. You re-enter the session with more intention. The mirror gets easier to look at because you have been looking at it regularly.
This is what AI makes possible that most tools don't: a daily, low-cost practice of self-observation in the act of working. Not journaling. Not reflection at a distance. A live question, asked at the moment the work closes, about who was actually in charge.
The machine doesn't care about the answer. But you should.
One session at a time, the Mirror Check builds something that no tool can build for you: the habit of knowing where you stand.
Close the session. Ask the question. Re-own what needs to be yours.
