
For you
Personal Practice
Module · What you can actually do with AI
The Personal AI Fluency Check
Fluency is not how many tools you have opened, it is what you can reliably get done with them. This check gives you an honest baseline across five parts of real competence: how broadly you can work, how deep you go on your main tool, whether you understand where these systems fail, whether you can tell good output from convincing output, and whether you can teach what you know. Answer for the habits you actually have, not the ones you mean to build.
What the five levels look like
Every dimension in this assessment is scored 1 to 5. This is what the levels mean, dimension by dimension. The graded report diagnoses where your own answers land and what to do about it.
Breadth across tasks
- 1One narrow use
- 2A couple of tasks
- 3Several, hit or miss
- 4A reliable range
- 5Wide and instinctive
At the low end: Using AI for one thing is a fine start and a low ceiling. Pick one task you do weekly that you have never tried it on, and run it through this week; breadth is built one task at a time. What good looks like: Reaching for AI instinctively across your work is real fluency. Watch that the habit does not outrun your judgement: the goal is knowing where it helps, not using it everywhere.
Depth on one tool
- 1Just ask questions
- 2Basic back and forth
- 3Some deliberate technique
- 4Work it deliberately
- 5Deep, near-automatic
At the low end: Asking questions is using ten percent of the tool. Spend one session learning what your main tool can actually do: how to give it context, how to shape its output; the return is immediate. What good looks like: Near-automatic depth on your main tool is where the real productivity lives. The risk now is missing that a different tool has quietly become better at your task; look up occasionally.
You know the failure modes
- 1Assume it is right
- 2Vaguely wary
- 3Know a few traps
- 4Know them concretely
- 5Anticipate them live
At the low end: Trusting AI by default is the single most expensive habit you can have with it. Learn the big three failure modes this week: fabrication, false confidence, and stale knowledge. They explain most bad outputs. What good looks like: Anticipating failure modes as you work is the mark of a genuinely fluent user. Keep the list current: each model release changes what the tool is good and bad at.
You can judge quality
- 1If it reads well
- 2Trust my gut
- 3Catch obvious errors
- 4Judge on substance
- 5Judge fast and reliably
At the low end: Judging output by how well it reads is exactly the trap fluent text is built to spring. Start checking one claim in every output against something you know is true; recalibrate your gut. What good looks like: Reliable, fast quality judgement is the skill that makes all the others safe. Guard it in the areas outside your expertise, where you have the least basis to catch a convincing mistake.
You can teach it
- 1Could not explain
- 2Show the buttons
- 3Explain some choices
- 4Teach the reasoning
- 5Others learn from me
At the low end: If you cannot explain how you use AI, your skill is a habit, not knowledge, and it will not survive a new tool. Try writing down your approach to one task; the gaps you find are where you do not yet understand it. What good looks like: When colleagues learn from how you work, your fluency has become an asset beyond yourself. Keep teaching the reasoning, not the recipe; the tools will change but the judgement carries over.