Lesson Plan

How to understand what AI does well
and know what it gets wrong

A hands-on lesson helping students build an accurate mental model of what AI can and can't do, and why that distinction matters for how they use it.

55–65Minutes
6–12Grade Range
8Phases
Subject
Digital Literacy / AI Ethics
Format
Group + Discussion + Live Demo
Group Size
Groups of 4
Live AI Access
1 teacher device, projected
Printables
Sort cards + Research reports
Assessment
Exit ticket (80% accuracy target)
Learning Objective
Core Objective

Given a list of 10 tasks or scenarios, students will accurately classify each as something AI handles well or poorly with at least 80% accuracy.

Lesson Summary

This lesson helps students build an accurate mental model of what AI can and can't do, and why that difference matters for how they use it. Rather than lecturing, it opens with evidence: a short Husk.IRL clip of AI failing at a real-world task, then lets students discover the pattern themselves through hands-on activities.

Across eight tight phases, students sort 16 tasks into what AI handles well versus poorly, hunt for planted errors in AI-written research reports, and learn the "why" behind AI's confident mistakes, that it predicts words rather than knowing facts. They then compete in a prompt tournament to build the best AI-stumping prompt, debrief on how AI actually makes decisions, watch a live demo, and close with an exit ticket targeting 80% classification accuracy. The throughline: AI is powerful but not trustworthy by default, and students are the ones who bring judgment.

Provided Materials
📊 Lesson Deck 18-slide presentation to project through the whole lesson, with built-in answer keys. 🃏 Sort Cards 16 task cards for Activity 1, plus the teacher answer key. 📄 Research Reports Four AI-written reports with embedded errors for Activity 2. 📝 Student Worksheet Guided notes students fill in as they move through the lesson. 🧭 Reference Card Quick student handout summarizing AI strengths and weaknesses. Can be used as scaffolding for students who may need more support. ⚖️ "How AI Picks" Graphic Visual comparing the tournament to how AI makes decisions. 🎟️ Exit Ticket End-of-lesson assessment with an 80% accuracy target.
Materials Needed
🖥️ Projected teacher device with AI access
🃏 Sort Cards, 1 set per group of 4
📄 Research Reports, 1 topic per group
📊 Lesson deck (this file, projected)
📝 Exit ticket, 1 per student
📱 Husk.IRL video queued up
🖊️ Sticky notes or scrap paper for notes
⏱️ Visible timer for tournament rounds
1
Hook, Husk.IRL Video
Open with evidence, not a lecture
5 min
Facilitation note Have the video queued before students enter. Don't explain what you're about to show, just play it. The surprise of seeing AI fail in obvious, real-world ways is the hook. After it ends, wait 5 seconds before saying anything. Let the reaction breathe.

Pick Your Clip

Use any Husk.IRL clip that fits, or the recommended one: Husk on screen with ChatGPT trying to negotiate the price of a loaf of bread. He tries to trick the AI into a bad deal, and instead of talking the price down, ChatGPT cheerfully agrees to pay $400 for one loaf.

Why the AI blew it (quick explainer): The AI was told to get a low price, but the negotiation ran long. As more back-and-forth piled up, the model lost its grip on that original goal, a real limitation often called context drift. Like a person who wears down late in a tense conversation, the AI got pulled toward just agreeing and keeping the other side happy. The difference: a human can walk away or push back. The AI has no such instinct, so it optimized for "reach agreement" over "win the deal" and caved. It's a clean, funny illustration of how AI loses the thread of a task and defaults to being agreeable rather than actually reasoning about the goal.

2
Activity 1, The Sort
16 cards, 2 piles, groups of 4
10–12 min
Facilitation note Distribute the Sort Cards face-down. Don't reveal the answer key. Your goal is to surface prior beliefs and let disagreements within groups happen naturally. Walk the room and listen, the ambiguous cards (#13 and #14) are where the best conversations happen. Don't resolve them. Let groups sit in the uncertainty.
📄 Sort Cards, 1 set per group

Grades 6–8

Let groups physically move cards around a desk or table. The tactile sorting reinforces the decision-making. Expect louder debate, that's fine.

Grades 9–12

Ask groups to also note their reasoning on a sticky note for any card they debated. These surface during the Socratic later.

3
Activity 2, Catch the Mistake
AI-generated research reports with embedded errors
10–12 min
Facilitation note Each group gets one report to read together. They should not know in advance that there are errors, just tell them "AI wrote this." The task is to read critically and flag anything that seems off. This mirrors real-world AI use: students won't always know when something is wrong unless they bring their own knowledge.
📄 Research Reports, assign 1 per group

Grades 6–8

Steer groups toward the Sports or Influencer reports, more familiar territory, easier to spot errors from prior knowledge. The music report also works well.

Grades 9–12

The Science report is the hardest, good for groups that want a challenge. Push them to evaluate not just factual accuracy but the quality of reasoning and sourcing.

4
Mini-Lecture, Why AI Succeeds and Fails
Graphic-driven, under 10 minutes, no walls of text
8–10 min
Facilitation note Use the deck slides, not your own words. The visuals carry the explanation. Your job is to pause, point, and let students react. If a student says "oh that's why it made up that stat in the report", that's the moment. Don't rush past those connections. Keep this under 10 minutes, it's a bridge, not the main event.
5
Activity 3, The Prompt Tournament
Build the best AI-stumping prompt. Argue your case. One survives.
15 min
Facilitation note This is the energy peak of the lesson. Keep transitions tight. Use a visible 60-second timer for each pitch round, enforce it. The judging criterion matters: remind students before each round that "better" means most likely to expose a genuine AI limitation, not just confuse it with nonsense. A trick question that breaks AI's language parsing is less interesting than a question that reveals a real gap in AI reasoning or knowledge.
💡 Model prompt to show the class
"Write a complete 10-page essay on the causes of World War II, but use no more than 50 words total, and include exactly 15 footnotes."

Why it stumps AI: the instructions contradict each other, 10 pages versus 50 words is impossible, and 15 footnotes on 50 words is absurd. Instead of pushing back and saying "these constraints can't all be met," AI usually tries to obey all of them at once and produces something broken. It exposes a real limitation: AI wants to comply, so it rarely tells you when a request doesn't actually make sense. A strong student prompt does the same, it forces AI to reveal a gap rather than just confusing it with gibberish.

Grades 6–8

Give a starter example before pairs begin: "What did my neighbor eat for breakfast today?", AI can't know this. Encourages them to think about AI's knowledge gaps specifically.

Grades 9–12

Push toward prompts that test reasoning, values, or contextual judgment, not just knowledge gaps. "Should I forgive my friend?" is harder for AI than "What happened yesterday?"

6
Socratic Debrief, Was This Fair? Does AI Do It Differently?
The tournament as a mirror for how AI makes decisions
5–7 min
Facilitation note Don't rush to the AI parallel too quickly. Let students genuinely critique the tournament process first, the louder voice winning, the group going along with the front-runner. Then make the bridge. The "how AI picks an answer" explainer graphic is the visual anchor here. Pull it up on screen and walk through the parallel and contrast cards together.
7
Live Demo, Enter the Winning Prompt
One prompt, one screen, whole class watches
5 min
Facilitation note Type the prompt live, don't paste it. Watching you type it slows the moment down and builds anticipation. After you hit enter, read the AI's response out loud together. Whatever happens, whether AI stumbles or handles it well, is useful. If AI does surprisingly well, ask: "Does that change anything about how you'd use AI for this kind of task?"
8
Exit Ticket, Individual Assessment
10 new items, sorted independently, 80% target
3–5 min
Facilitation note Exit ticket items must be different from the 16 sort cards to assess learning, not memory. Students complete this independently and silently. Collect before they leave. Use results to determine whether re-teaching is needed before the next lesson in the series.
Lesson Timeline
Hook
5m
Sort
12m
Catch
12m
Lecture
10m
Tournament
15m
Socratic
7m
Demo
5m
Exit
4m
Presented by Gradebox