Use it, judge it, explain it

This is the project the whole subject builds toward. You will pick a real task from your own life or study, use AI to help complete it, and document the entire process honestly: every prompt, every decision, every check. The finished task matters, but the documentation matters more, because it is where you prove you understand what the tool was actually doing and where you were the one doing the thinking.

Choose a task

• A study task: turning your class notes on a topic into a revision guide, then checking every claim in it against the textbook.
• A writing task: drafting something real, like a job application, a letter or an article for the school newsletter, through several rounds of iteration.
• A research task: getting your head around a topic you genuinely need to understand, with every factual claim verified against a source the AI did not write.
• A practical task: planning something with real constraints, like a trip, an event or a training program, where you can test whether the plan survives contact with reality.

What your documentation must show

Keep the log as you go, not from memory afterwards. Record each prompt, why you wrote it that way, and what came back. When an output was wrong or useless, keep it in the log and note what you changed, because the dead ends are evidence of real work rather than something to hide. Verify the claims that matter against sources you found yourself, and record exactly what you checked and what you found. Then write the honest page: where the AI genuinely saved you time, where it failed or misled you, and which parts of the finished task only existed because a person was steering.

Good projects are not the ones where the AI performed flawlessly. They are the ones where the student can explain, in their own words, why the tool behaved the way it did and what they would do differently next time.