Talk it through with Aurelius
Library›Aurelius›The problem
Aurelius · Work & Leadership
Knowledge + Guidance

Can I trust the AI I use to measure my work?

Name it plainly: you hand your numbers to a machine, then wonder if you should believe what comes back. That is not weakness. That is what happens when you use a tool you cannot open up and inspect. But look closer. The fear is not really about AI. It is about you not knowing where the tool's sight ends and its guessing begins. You have made yourself a passenger in your own judgment. That is the part within your power to fix. The work is not to trust the tool more or less. The work is to find its edges yourself — where it is sharp, where it is dull — and choose, each time, how much weight its answer deserves.

◆ How this problem reads on the two dials
GuidanceKnowledge
More coaching
Some to learn
1:1 with AureliusWith others (a Pod)
Mostly you & the coach
A little with peers
The person needs a few clear ideas about how AI assumptions work, but mostly needs a repeatable habit of testing and checking, which is coaching territory.
How the two dials adapt to you →
What’s really going on

Trust is the wrong question. Ask instead: where does this tool see clearly, and where is it blind? Pick one number it gave you. Check it against something you know is true. That single test tells you more than any amount of worrying will.

🔒 What you’ll build togetherUnlock by starting
A moveTake one number the AI gave you this week. Check it by hand, or against a source you already trust. Write down how far off it was.
A moveList what the tool cannot see about your exact situation — what it was trained on, what data it lacks. Keep this list where you'll see it again.
A moveBefore you act on any AI output, ask yourself: what would I do differently if this number were wrong? If you have no answer, you are not ready to use it.
A moveKeep a short, running record of every time the AI has been wrong before. This is how you learn its blind spots, not by guessing.
A moveSet yourself one rule: no important decision rests on an AI number alone. Always pair it with one number a human has checked.

What changes unlock by starting

  • You know, specifically, where this tool is reliable and where it is not — not in general, but for your actual work.
  • You stop receiving AI output as fact and start treating it as a claim you check.
  • You make decisions with your eyes open, on evidence you gathered, not on quiet faith.
  • You can explain to anyone — a boss, yourself — exactly why you trust a given number, or why you don't.

Related problems

One object, two jobs: a public answer to a real problem, and — the moment you start the chat — Aurelius’s live plan for your version of it.