Lesson 1, Topic 1
In Progress
Unit 7 – Teacher Resources
Instructor Resource · AIBIP Course
Unit 7 Teacher Resources
AI and Problem Solving
Everything you need to teach, assess, and support students through Unit 7 — lesson guides, learning objectives, discussion prompts, common misconceptions, differentiation strategies, a pacing guide, and the full project rubric.
Why Unit 7 Matters
Unit 7 is where the AIBIP course turns knowledge into skill. Earlier units taught students what AI is, how it learns, and where it's used. This unit asks the practical question: how do you actually solve a problem with AI? Students learn that AI's power comes from analyzing large amounts of data, that its real role is to amplify human problem-solving rather than replace it, and that the same handful of methods reappear across every domain.
The through-line is a five-step framework — Define → Data → Build → Test → Iterate — that mirrors industry processes like CRISP-DM. The most common instructor pitfall is letting students rush past the "Define" step to jump into a tool or model. Mastery means framing a problem clearly, choosing an appropriate approach, and checking results with critical thinking. That combination of structure and judgment is exactly what the Unit 7 project assesses.
🎯 Teaching through-line: Keep returning to real, nameable examples (navigation apps smoothing traffic, DeepMind cutting cooling energy 40%, UPS's ORION saving millions of miles, the city water-leak case study). Push students to connect each example to a framework step and a recurring method — that's what makes the learning transferable.
Unit 7 At a Glance
3
Lessons: 7.1 through 7.3
2.5–3 hrs
Estimated student completion time
12
Knowledge-check questions across lessons
12 pts
Solve a Real-World Problem with AI project
| Lesson | Title | Core Concepts |
|---|---|---|
| 7.1 | AI for Real-World Problem Solving | AI as a problem-solving partner; data analysis; personal / community / global scales; the human + AI loop; critical thinking |
| 7.2 | AI Solutions in Various Domains | Healthcare, environment & logistics; recurring methods (classification, prediction, clustering, anomaly detection); benefits and limits |
| 7.3 | Frameworks for AI Problem Solving | Five-step framework (Define, Data, Build, Test, Iterate); problem scoping; prototype, evaluation, iteration, deployment; CRISP-DM |
Lesson-by-Lesson Teaching Guides
Lesson 7.1 — AI for Real-World Problem Solving
Students will be able to
- Recognize how AI helps solve everyday problems and complex global challenges by analyzing data
- Explain the role of critical thinking in deciding if AI is the right tool
- Give examples of AI solving problems at personal, community, and global scales
- Practice using an AI assistant to brainstorm solutions to a real-world problem
Key teaching points
- Stress that AI amplifies human problem-solving — it brings speed and pattern-spotting; humans bring judgment and context.
- Use the human + AI partnership loop: humans frame, AI analyzes, humans judge and act.
- Emphasise that a prediction (like a flood warning) is only useful if humans have a plan to act on it.
Activity note
- The Problem-Solving Lab is a simulated chatbot — students practise prompt engineering. Have them compare a vague prompt with one that adds a constraint ("inexpensive, easy for a student") and notice how answers sharpen.
Lesson 7.2 — AI Solutions in Various Domains
Students will be able to
- Identify how AI solves problems in healthcare, environmental sustainability, and logistics
- Describe concrete examples and their impact in each domain
- Explain a real benefit and a real limitation of AI in each field
- Match a real-world problem to the AI method best suited to solve it
Key teaching points
- Draw out the recurring methods: classification, prediction/regression, clustering, anomaly detection. The context changes; the method reappears.
- Anchor each domain on one memorable stat: 2× stroke-scan accuracy, DeepMind's 40% cooling cut, UPS's millions of miles saved.
- Always pair a benefit with a limitation (privacy, data quality, changing conditions) — this builds realistic expectations.
Activity note
- The Domain Match Lab has students match problems to methods via dropdowns. Discuss why each fits — the "shape" of a problem (labeling vs. forecasting vs. grouping) points to the method.
Lesson 7.3 — Frameworks for AI Problem Solving
Students will be able to
- Apply a five-step framework: Define, Data, Build, Test, Iterate
- Explain the key question and critical-thinking check at each stage
- Understand why an AI framework loops back rather than running in a straight line
- Sketch a simple, well-scoped AI project plan of their own
Key teaching points
- Spend real time on problem scoping — the 4Ws and measurable success criteria. This is where projects succeed or fail.
- Reinforce "garbage in, garbage out": weak data or a fuzzy problem produces a weak solution.
- Frame iteration as normal and expected — the first model is rarely the last. Connect to CRISP-DM as the industry version.
Activity note
- The Framework Sequencer has students order the five steps for a cafeteria-prediction scenario. Use the reflection questions to discuss why "Define" must precede "Data."
Discussion Prompts
Lesson 7.1
What's a problem in your own life or community that AI could help with — and what data would it need?
Draw out that AI needs data to learn from. Reinforce that AI supports, not replaces, human judgment about what to do with the answer.
Lesson 7.1
Why isn't an AI prediction, by itself, a complete solution?
Use the flood-warning example: a prediction only helps if people have a plan to act on it. Humans stay in the loop to interpret and respond.
Lesson 7.2
How can the same AI method solve problems in totally different fields?
Connect image classification (reading X-rays / spotting diseased crops) and regression (energy use / product demand). The method transfers; the data changes.
Lesson 7.2
When might a technically "accurate" AI still fail to solve the real problem?
Explore data quality, changing conditions, and trust. A route optimized only for distance may hit traffic; a model must serve the real goal, not just a metric.
Lesson 7.3
Why is "define the problem" more important than jumping straight to a model?
Highlight that a fuzzy problem leads to solving the wrong thing. A well-scoped problem with clear success criteria guides every later step.
Common Misconceptions & Fixes
Misconception
"If you have an AI tool, you can skip planning and just feed it data."
The Fix
Without defining the problem first, you may solve the wrong thing. The framework starts with scoping, not tooling.
Misconception
"AI replaces human decision-making in problem solving."
The Fix
AI amplifies humans. It provides data-driven insights; humans frame the problem, judge the output, and act on it.
Misconception
"A model that scores perfectly on its training data is the best model."
The Fix
That can be memorization. Evaluate on unseen test data — the goal is generalizing to solve the real problem.
Misconception
"If the first AI attempt fails, the project should be abandoned."
The Fix
AI projects iterate. Collecting more data, tweaking the model, or re-scoping the problem is a normal part of the process.
Misconception
"A fancy AI model is always better than a simple rule."
The Fix
Sometimes a simple rule solves the problem nearly as well and is far easier to deploy. Choose the approach that fits.
Misconception
"Data quality doesn't matter much if the algorithm is good."
The Fix
"Garbage in, garbage out." Even the best algorithm produces poor results from messy, biased, or missing data.
Differentiation Strategies
🟢 Support
Provide a printed glossary of the Unit 7 framework terms to reference during activities
Pair students to complete the Problem-Solving Lab and Framework Sequencer together
Offer a fill-in-the-blank project template with each framework step pre-labeled
Model one project example (e.g. the cafeteria scenario) together as a class
🔵 Core
Complete each lesson, activity, and knowledge check independently
Plan an original problem through all five framework steps in the project
Use the self-assessment rubric to reflect on their own work
Explain the Define → Data → Build → Test → Iterate loop in their own words
🟣 Extend
Actually build and evaluate a working prototype (e.g. a scikit-learn classifier)
Compare two AI approaches to the same problem and justify the better choice
Research a real CRISP-DM or industry case study and map it to the five steps
Add a data-ethics analysis: what biases or privacy risks does their solution raise?
Unit 7 Project Rubric — Solve a Real-World Problem with AI (12 pts)
Students identify a problem, propose an AI solution, and plan it through the framework — defining the problem, the data needed, the steps, and the expected impact. Each criterion is worth up to 3 points (Excellent = 3, Satisfactory = 2, Needs Improvement = 1), for 12 points total. A score of 10/12 or above is a strong pass.
| Criterion | Excellent (3) | Satisfactory (2) | Needs Improvement (1) |
|---|---|---|---|
| Problem Definition & Planning | Clear, insightful problem, rationale, and success criteria; logical plan addressing all key steps | Clear problem and goal; plan covers major steps, some detail refined along the way | Vague or missing problem; little evidence of planning |
| Technical Implementation | Solution implemented correctly and efficiently; AI technique applied well, maybe with extras | Functional solution that achieves the task; competent use of tools, minor rough edges | Incomplete or error-prone; tools misused; heavy uncredited copying |
| AI Solution & Analysis | Well-justified approach; thorough, correct evaluation with insightful interpretation and next steps | Appropriate approach; at least one metric with a reasonable interpretation tied to the goal | No real AI component or unsuitable approach; minimal or incorrect evaluation |
| Documentation & Presentation | Clear explanations at each stage; well-structured; effective visuals; thoughtful reflection | Explanations for most steps; mostly understandable; a conclusion or reflection provided | Sparse or hard to follow; little narrative; messy formatting |
✅ Grading tip: The project's on-page self-assessment lets students score themselves before submitting. Use their self-scores as a conversation starter — where they over- or under-rate themselves reveals a lot about their understanding of the framework.
Suggested Pacing
| Session | Focus | Time |
|---|---|---|
| Session 1 | Lesson 7.1 — AI as a problem-solving partner, scales of problems, Problem-Solving Lab, knowledge check | 45–60 min |
| Session 2 | Lesson 7.2 — AI across healthcare, environment & logistics, Domain Match Lab, knowledge check | 45–60 min |
| Session 3 | Lesson 7.3 — the five-step framework, water-leak case study, Framework Sequencer, knowledge check | 45–60 min |
| Session 4 | Unit 7 Project — Solve a Real-World Problem with AI + self-assessment | 60–90 min |
📌 Flexible delivery: Sessions map to class periods but can be split or combined. For a compressed schedule, assign the knowledge checks and project as homework. For an extended schedule, use the "Extend" ideas to add a build-and-evaluate day where students implement a real prototype.
Instructor FAQ
To understand the framework, no — the lessons focus on how to think about AI problems. For the project, coding from Unit 6 helps if students build a working prototype, but a strong, well-reasoned plan can earn full marks on its own. Emphasise clear thinking over clever code.
No. The Problem-Solving Lab, Domain Match Lab, and Framework Sequencer are self-contained simulations built into the lessons, so every student can complete them without an account or external tool. A real AI assistant makes a great optional extension.
Each lesson's knowledge check is a short, self-grading quiz that gives instant feedback and highlights the correct answer. They're formative — use them to surface misconceptions (especially around problem scoping and evaluation) rather than as high-stakes grades.
Use the "Extend" differentiation ideas: build and evaluate a real prototype, compare two AI approaches, map a CRISP-DM case study to the five steps, or add a data-ethics analysis. The project's open-ended options (health, environment, logistics, or their own idea) also scale up naturally.
Unit 7 is where students apply everything: the AI types from Unit 1, the ethics awareness from Unit 4, and the data and coding skills from Unit 6 — all in service of solving a real problem. The framework becomes a reusable checklist for any AI work that follows.