Use a structured interview to build a useful context file. Follow Carla's demonstration, pause to make your own attempt, and compare your check with the result.
Your activity / Right here
Bring a notebook. Begin below.
The activity is included on this page. Work on paper or in your own document; keep your first attempt and the evidence you checked.
AI Lab: My AI Interviews Me
Use only an AI tool your instructor approves and that allows users your age. Check its minimum age and parent-permission rules first; an adult's permission cannot override an 18+ limit. Follow your school's rules, and do not share personal information. About 30 minutes.
1. Type your handwritten context file into a plain text or notes file named my_context.md.
2. Paste this prompt into your approved AI tool: "Interview me to improve my context file. Ask one question at a time, and no more than eight questions. After each answer, suggest one line for the file and label it SUGGESTED. Do not add anything I did not say. Do not ask for my name, school, address, or any personal information."
3. Decide on every suggested line in a table like this one:
| Suggested line | Accept, Edit, or Reject? | Why |
|---|
| | |
4. Build one class context file the same way. Choose a class, name the file (for example, algebra.md or biology.md), and include: Current Unit, What I Already Know, Mistakes I Make, How My Teacher Wants Work Shown, and AI Rules for This Class.
5. Verify: Read both files aloud. Check that every line is true, that every line is something you actually said, and that every line passes Needed, Mine, and Safe. Check the class AI rules against your instructor's official policy, not your memory. Delete any line the AI added that you did not approve.
6. Record your AI use: which tool you used, what you asked, and which suggested lines you accepted, edited, or rejected.
Own It
No AI. About 10 minutes.
Print or copy out my_context.md V1. Read it as if you were a substitute teacher meeting you for the first time. Would it help them help you? Circle one line to improve next month.
Then finish this sentence: "The line in my file that protects my learning the most is __________ because __________."
Portfolio piece: my_context.md V1 and one class context file.
Context File Upgrade
Assemble my_context.md V1. Put all your sections in one file, in this order: About Me, My Core Values, How I Think First, What I Believe About AI So Far, How to Give Me Context, My News Rules, Verify Rules for My AI, Mistakes to Watch For, Never Paste, Privacy Rules for My AI, My Week, and Rules for My AI. Put the version and date on the first line:
# My Context (V1, updated YYYY-MM-DD)
Review the file once a month. When something changes, update it and raise the version number: V2, V3, and so on.
Skeptic's Corner: What Does the Study Actually Say?
These statements are exercises to classify and check. Do not assume every statement is true.
Use the six labels from Chapter 1: Fact, Idea, Rob's Story, Made-Up Example, Rule, Prediction. Then catch the weak claims.
1. "Markdown was introduced on March 15, 2004."
2. "Mike's half-page file helped more than his two-page file."
3. "A classmate says: 'That study proves AI makes students worse at math.'"
4. "Giving your AI permission to admit uncertainty can drastically reduce false information."
5. "Use only an AI tool your instructor approves."
6. "Soon, AI will know you so well that you will never need a context file."
Read the video transcript
Read along, pause, or return to an idea. This transcript follows the selected lesson video.
I'm Carla, an AI avatar.
Today I'm going to let an AI interview me to improve my
chemistry context file.
The AI asks.
I answer.
And I approve every single line before it goes in.
Let's go.
I always start on paper.
Here's my draft: my current unit,
what I already know, the mistakes I make,
and how my teacher wants work shown.
It's short and a little messy.
That's fine.
The AI will help me improve it, but it starts as mine.
Why chemistry?
Because a class file stays with one class.
My personal file, my_context.md,
goes everywhere.
This file only has to know what chemistry needs right now.
Now the prompt.
I set the rules before the interview starts.
Interview me to improve my context file for chemistry class.
Ask one question at a time.
After each answer, suggest one line for the file and label
it suggested.
Do not add anything I did not say.
Do not ask for personal information like my name,
school, or address.
First question: what part of chemistry is hardest for you right now?
I answer: converting grams to moles.
Suggested line: I find gram to mole conversions difficult.
That's true, and it's what I said.
Accept.
Next: how does your teacher want work shown?
I say: units on every number, and the conversion factor
written out.
The suggested line matches, so I accept it with one edit.
I add, including units in the final answer.
Another question: how do you like to study?
I say: practice problems first, then explanations.
Suggested line: quiz me before you explain.
That one goes straight into my rules.
Accept.
Now watch this one.
The AI suggests: Carla struggles with math in general.
I never said that.
That's the AI filling a gap with a guess.
Reject.
And this one: what school do you go to,
so I can match your curriculum?
Nope.
That fails my Safe test.
I answer: I won't share that.
It's a standard high school chemistry course.
Here's the thing.
I told it not to ask, and it asked anyway.
AI tools can miss an instruction.
That's exactly why I'm the one approving lines.
I keep a table as I go.
Every suggested line gets accept,
edit, or reject, and a reason.
By the end, I have eight decisions,
and every one of them is mine.
Notice what the interview is good for.
It finds gaps I didn't think about,
like how my teacher wants work shown.
But the answers, and the decisions,
stay with me.
Before I save, I run the privacy line on the whole file.
Needed?
Mine?
Safe?
One line mentions my lab partner's name.
That's her information, not mine, so it goes.
I update the top line: version one,
with today's date.
Then I verify.
I read the file out loud.
Is every line true?
Did I actually say it?
I also check the class AI rules against my teacher's
written policy, not my memory.
Last, I record my AI use: which tool,
what I asked, and which lines I accepted,
edited, or rejected.
That note is how I stay honest with myself and my teacher.
The AI asked the questions.
I wrote the answers.
Every line in this file is one I approved.
Your turn: open the AI Lab and let your AI interview you.
Before you use an AI tool
Follow your course or institution's rules. Use an instructor-approved tool that permits your age, and meet any required consent conditions. Keep names, contact details, school records, and other private information out of prompts. Use fictional examples when you can.
Learning without an AI account
Follow the demonstration, pause before each answer, and do the reasoning on paper. Discuss what you would ask, how you would check it, and which decisions would remain yours.
About the characters and voices
This course uses AI-created characters and synthetic voices. Carla is an AI avatar. They help tell the story; they are not real students sharing testimonials.
Designed for high-school and community-college learners. Choose examples from your classes, work, and everyday life. Keep your completed work in your course portfolio.
