Week 6 · Lesson 1 · 50-60 minutes

Choose the Right Problem and Tool

Start a design project the way real teams do: name a user need, write a clear problem definition, and decide whether AI is even the right tool for the job.

Materials: This lesson in a web browser · Paper and pencil, or a notes app
This lesson: In progress
Auto-saves on this device

What you'll be able to do

Turn a user need into a clear, one-sentence problem definition.
Judge AI suitability: decide whether a problem is a good fit for AI or for fixed rules.
Name the input and output your project would use.
Explain why choosing the right problem matters before building anything.

Think about it

A school club says, 'We want to use AI!' But for what? A design team never starts with the tool — it starts with a person who has a problem. Whose problem would your project solve, and what exactly is it?

This week you plan a real project. Everything starts with the problem, not the technology.

Make a prediction

Predict: of these two ideas — 'sort library book photos into fiction and non-fiction' and 'unlock a door with the correct password' — which one is a better fit for AI?

Vocabulary

See the full course vocabulary

User need: A real problem a specific person or group has that they want solved, described from their point of view.
Problem definition: A clear, one-sentence statement of exactly what you are trying to solve and for whom.
AI suitability: How well a problem fits AI, which learns patterns from many examples, versus a fixed rule a person can just write.
Input: The information the system would take in, like a photo, a sentence, or a set of numbers.
Output: The result the system would give back, like a label, a prediction, or a suggestion.

Start with a user need, not a tool

Good projects begin with a real person who has a real problem — a user need. 'Kids in my class forget which recycling bin to use' is a user need. 'I want to use AI' is not; it names a tool, not a problem.

When you start from a user need, you can tell later whether your project actually helped. You describe the need from the user's point of view: who they are, and what is hard for them right now.

A gardener can't tell which leaves show plant disease
New students get lost finding classrooms on their first day
A pen-pal club gets messages in languages members don't read

Write a clear problem definition

Once you have a user need, you write a problem definition: one clear sentence that says exactly what you are solving and for whom. A sharp definition keeps a project from drifting. 'Help sixth graders sort classroom waste into recycling, compost, and trash from a photo' is specific enough to build and to test.

A vague definition like 'make recycling better' can't be tested — you would never know if you succeeded. A good problem definition names the user, the task, and what a good result looks like.

Vague: 'Help with plants.' Clear: 'Tell a gardener whether a tomato leaf photo looks healthy or diseased.'
Vague: 'Make school easier.' Clear: 'Give a new student directions to a room from its number.'

Decide AI suitability: is AI even the right tool?

Not every problem needs AI. AI shines when a task is too messy to write exact rules for but you have many examples to learn from — like telling healthy leaves from diseased ones in photos. A problem that one clear rule already solves does not need AI: 'unlock if the password matches' is a rule, not a learning task.

To judge AI suitability, ask three questions: Is the task too messy for hand-written rules? Could you collect many labeled examples? And is it okay if the system is sometimes wrong? If the answers are yes, AI may fit. If a simple rule works, use the rule — it is cheaper, clearer, and easier to check.

Good fit for AI: sorting photos of waste into recycling, compost, or trash.
Bad fit for AI: turning on the hallway lights at 7:00 a.m. — that's a fixed rule.
Bad fit for AI: a task where being wrong even once could seriously hurt someone.

Worked example

From a fuzzy wish to a buildable project

Start with the wish: 'I want to use AI to help the recycling club.'
Find the user need: club members and other students often put waste in the wrong bin because they can't tell what's recyclable.
Write the problem definition: 'Help students decide which of three bins — recycling, compost, or trash — an item belongs in, from a photo of the item.'
Name input and output: Input is a photo of one item; Output is a label — 'recycling', 'compost', or 'trash'.
Check AI suitability: Too messy for exact rules? Yes, items vary a lot. Can you collect many labeled photos? Yes. Okay to be sometimes wrong (a person still checks)? Yes. AI fits.

Takeaway: A buildable project has a real user need, a one-sentence problem definition, a named input and output, and a yes on AI suitability.

Should this problem use AI?

A decision tree for AI suitability. Start: 'Can one clear rule solve it?' If yes, use a rule, not AI. If no, ask 'Can you collect many labeled examples?' If no, AI probably won't work yet. If yes, ask 'Is it okay if it's sometimes wrong, with a person checking?' If no, be very cautious or don't use AI. If yes, AI may be a good fit.

Use a rule when a rule works; save AI for messy tasks with examples and human checking.

Problem definition checklist

Part	Question it answers	Weak example	Strong example
User	Who has this problem?	'People'	'New sixth-grade students'
Task	What exactly are we solving?	'Help with directions'	'Give a route to a room from its number'
Input	What goes in?	'Some info'	'A room number'
Output	What comes out?	'An answer'	'Step-by-step directions'

A table of what a strong problem definition includes. User: names a specific person or group. Task: says exactly what to decide or produce. Input: names what the system takes in. Output: names what it gives back. Testable: you could tell later whether it worked.

Activity

Problem-Tool Fit Studio

Activity briefing

Choose a project problem, write its problem definition, and decide whether AI is the right tool.

Read the candidate problems and choose one that interests you, or write your own user need.
Write the user need from the user's point of view: who they are and what is hard.
Write a one-sentence problem definition naming the user, the task, and what a good result looks like.
Name the input and the output your system would use.
Run the AI-suitability check and write one sentence: does AI fit, or would a simple rule do?

You will review a set of candidate problems, pick one (or bring your own user need), and turn it into a buildable project brief: a user need, a one-sentence problem definition, a named input and output, and a decision about AI suitability. This is the seed of your Final Project.

You will need

Paper and pencil, or a notes app

Success looks like

A user need written from the user's point of view.
A one-sentence problem definition that names the user, task, and a testable result.
A named input and output.
A clear AI-suitability decision with a one-sentence reason.

Built-in dataset

Candidate project problems

A built-in set of starter problems for the studio: sort classroom waste from a photo, flag plant-disease leaves, group lost-and-found items, translate pen-pal messages, sort library books by topic, and 'unlock a locker with a password' (a rules-only trap to reject). Each card lists a rough user, a possible input, and a possible output.

For now, work through this activity with the built-in materials, your notes, and discussion. An interactive version is coming in a later update.

Check your understanding

Check that you can define a problem and judge whether AI fits it.

Question 1. Which is the strongest problem definition?
Question 2. Decide if the statement is true or false.
If one clear rule already solves a problem, you should still use AI because AI is more advanced.
Question 3. Which project is the best fit for AI?
A club is choosing between: (A) ring a bell exactly when class ends, or (B) tell from a photo whether a plant leaf looks healthy or diseased.

0 of 3 answered

Try it yourself

Three problems, one keeper

Come up with three user needs, then pick the one that best fits AI and defend your choice.

Write three different user needs, each from a real user's point of view.
For each, name a possible input and output.
Run the AI-suitability check on all three and choose the best fit.
Write two sentences defending why your keeper fits AI better than the others.

Success looks like

Three user needs, each with an input and output.
An AI-suitability decision for each.
A chosen problem with a two-sentence defense.

Reflection

Saved on this device as you type

Please don't include private information. Your answers are saved only in this browser and are never sent anywhere.

Why is it tempting to start with the tool ('let's use AI') instead of the problem, and what goes wrong when you do?

Which part of writing a problem definition was hardest: naming the user, the task, or the testable result?

Recap

Great AI projects start from a real user need, a clear problem definition, and an honest check of whether AI is the right tool at all.

A user need describes a real person's problem, not a tool you want to use.
A problem definition names the user, the task, the input, the output, and a testable result.
Judge AI suitability: use a rule when a rule works; save AI for messy tasks with many examples.

Grades 7–8 extension

Opportunity or hype?

Companies sometimes add 'AI' to a product mostly for marketing, even when a simple rule would work better. That can waste money and make a product harder to trust and check.

Find a real product or app that advertises an 'AI' feature. Decide whether the task truly needs AI or whether a fixed rule would do. Write a short paragraph explaining your judgment.

Finish this lesson

Attempt the knowledge check — not done
Save a reflection (recommended) — not done

Attempt the knowledge check to finish this lesson.