Final project: The chatbot survival guide
Applies all four objectives (S1.1–S4.1). Students create a guide for a younger student (or a grandparent — their choice of “someone who trusts chatbots too much”) explaining how to use a chatbot wisely. Scored with the rubric (school adaptation noted there) and planned with the project planner.
Last reviewed: 2026-10-25
The brief (student-facing)
A younger student says: “I just ask the chatbot — it knows everything.” You have four lessons of evidence that this is not true. Make them a survival guide.
Format: a poster, a one-page leaflet, a short video script, or a web page. Paper formats are fine; no account with any AI service is needed.
It must include:
- How it works — in your own words: what the chatbot actually does (Lesson 1). Include one example of a sentence where the next word is predictable, and one where it isn’t.
- When it lies beautifully — the claim types that must be checked (numbers, dates, names, sources, superlatives) and one real fact-check you did yourself, with your source (Lesson 2).
- Who it might leave out — one way outputs can be unfair to a group and what causes it (Lesson 3).
- Your rules — two personal rules for using AI for schoolwork, with reasons (Lesson 4).
Constraints: 200–300 words (or 90 seconds spoken). Honest tone — no scare tactics; the guide should make the reader smarter, not scared.
Cite your AI help: end with one line saying whether — and how — you used AI while making the guide.
How the project is scored
Mapped to the rubric:
| Rubric criterion | In this project |
|---|---|
| Accuracy checking | Part 2 — the self-done fact-check with a real source is the core evidence |
| Data judgment | Adapted: source judgment — the guide cites where its own facts came from |
| Bias recognition | Part 3 — example and cause must match |
| Justification | Parts 1 and 4 — explanations in the student’s own words, reasons for rules |
What the score does not measure: design beauty. A plain, correct guide outscores a gorgeous, vague one. Say this out loud when you assign it.
Suggested timing
| Session | Work |
|---|---|
| Lesson 4, end | Brief + planner started (choice of format, partner choice) |
| Homework or one work period | Draft all four parts; do the fact-check |
| Final 20 minutes (next lesson) | Peer review in pairs against the four rubric criteria, then revise |
Peer review protocol (the final 20 minutes)
Pairs swap guides and answer three questions on a sticky note:
- “Which part convinced you most — why?” (names a specific part)
- “Where did you want a source or example and not find one?”
- “Is any part scare-tactics instead of explanation? Quote it.”
Authors revise using the notes; the teacher scores after revision.
Common pitfalls to warn about
- Scare-posters (“AI WILL RUIN YOUR HOMEWORK”) — fails the tone constraint; the brief asks for smarter, not scared.
- The fact-check that checked nothing — “I checked it and it was true” without naming the source scores low on criterion 1.
- Rules without reasons — “don’t use AI for homework, because it’s bad” is not a justification; push for the why.
- Copying the lessons back verbatim — the guide must use the student’s own words; that is what “in your own words” is testing.