STEM with AI: From Question to Discovery

Table of Contents

1. Overview
2. How to Use This Framework
3. What You'll Learn
4. Understanding AI Tools
5. AI Is Your Partner, Not Your Replacement
6. The 8 Steps at a Glance
7. Step 1: Define Your Project
8. Step 2: Set Up Your Tools
9. Step 3: Design Your Application
10. Step 4: Choose Your Tech Stack
11. Step 5: Design UI/UX
12. Step 6: Build & Test Your Application
13. Step 7: Collect Data & Observations
14. Step 8: Analyze Data & Draw Conclusions
15. Project Completion Checklist
16. Glossary
17. Tips for Working with AI
18. Resources

Overview

This framework combines the scientific method with software development, providing a structure from which students can devise their own STEM project. Within that project, students develop an application that helps them investigate a real-world question, collect data, and draw conclusions—all with the help of AI coding tools.

Who is this for?

• Middle school students (Grades 6-9)
• Students with little or no prior coding experience
• Science fair participants who want to incorporate technology
• Teachers and mentors guiding student projects

What makes this different? This framework is AI-first. Students use AI assistants throughout every phase—not just for coding, but for brainstorming, designing, debugging, and analyzing results. The student's role is to lead the project, ask good questions, and make decisions while AI handles the technical heavy lifting.

How to Use This Framework

Work through the 8 phases in order. Each phase includes:

• Explanations of what you're doing and why
• Prompts to try — copy these into your AI assistant to get help
• Tables and checklists — fill these in as you go to document your project

Your job isn't to do everything yourself. Your job is to:

• Ask good questions
• Give clear instructions to AI
• Check that AI's work makes sense
• Make decisions about your project
• Understand what you're building and why

Think of AI like a very smart assistant who can do almost anything—but needs YOU to be the project leader.

What You'll Learn

By completing a project using this framework, you'll gain experience in:

These aren't just school skills—they're the same skills used by scientists, engineers, and developers in the real world.

Understanding AI Tools

Before you start your project, let's talk about the AI tools available to help you. There are two main categories: AI assistants (for conversations and help) and AI coding tools (for building software).

AI Assistants

These are conversational AI tools you can chat with to brainstorm ideas, get explanations, and work through problems.

For this framework: Any of these AI assistants will work well. The free tiers are sufficient for most student projects. Pick whichever one you or your school already has access to.

AI Coding Tools

These are specialized tools that help you write code. They combine AI assistants with code editors.

Choosing Based on Your Budget

Your Friend Markdown

As you work with AI, you'll notice it often writes things in a format called Markdown. This is a simple way to format text that's easy to read and write.

Why should you care?

• AI naturally writes in Markdown, so understanding it helps you read AI's output
• It makes your notes, documentation, and project write-ups look professional
• GitHub, Google Docs, Notion, and many other tools understand Markdown
• It's much faster than clicking formatting buttons

Basic Markdown you'll see:

You don't need to memorize Markdown—AI will write it for you. But recognizing it helps you understand what AI gives you and make quick edits yourself.

AI Is Your Partner, Not Your Replacement

This is the most important thing to understand: AI is here to help you do YOUR project, not to do the project for you.

Think of AI like having a super-smart study partner who:

• Never gets tired of answering questions
• Can explain things in different ways until you understand
• Types really fast and knows a lot about coding
• Is happy to do tedious work while you focus on the interesting parts

But here's what AI cannot do:

• Have your curiosity or come up with questions that matter to YOU
• Make decisions about YOUR project
• Know if something is actually working without you testing it
• Present your project or explain what you learned
• Feel proud of what you built

What AI Does for You

The "Grunt Work" vs. The Real Work

There are two kinds of work in any project:

A Good Way to Think About It

You are the scientist. AI is your lab assistant.

You are the architect. AI is your construction crew.

You are the chef. AI helps prep ingredients and wash dishes.

The final product has YOUR name on it because YOU made the important decisions.

How to Know If You're Using AI Right

Ask yourself these questions:

• ☐Can I explain what my project does and why?
• ☐Do I understand the main parts of my code (even if I couldn't write it from scratch)?
• ☐Did I make the key decisions, or did I just accept whatever AI suggested?
• ☐Could I answer questions about my project from a teacher or judge?
• ☐Do I know what my data means and why my conclusions make sense?

If you can answer "yes" to these questions, you're using AI the right way—as a powerful tool that helps YOU accomplish YOUR goals.

The 8 Steps at a Glance

Step 1: Define Your Project

In this phase, you'll establish the scientific foundation for your project. You'll work through the key elements of the scientific method:

Working with AI to Find Your Topic

Start a conversation with your AI assistant:

Building Your Research Question

A good research question can be measured and tested.

Building Your Hypothesis

Template:

If [I change this variable], then [this will happen], because [scientific reasoning].

Identifying Your Variables

Step 2: Set Up Your Tools

Before you start building, you need to gather all the tools required for your project. This includes tools for your experiment and tools for development.

Identify Your Project Tools

Experiment Tools

Development Tools

Setting Up Your Tools

Step 3: Design Your Application

Describing Your App to AI

The better you describe what you need, the better AI can help you build it.

User Stories

User stories describe WHO does WHAT and WHY.

Feature List

Must Have (MVP):

• ☐[Feature 1]
• ☐[Feature 2]
• ☐[Feature 3]

Nice to Have (Later):

• ☐[Feature 4]
• ☐[Feature 5]

Step 4: Choose Your Tech Stack

Letting AI Recommend Technologies

You don't need to know all the technology options—AI does.

Understanding AI's Recommendations

My Tech Stack

Step 5: Design UI/UX

Creating Your App's Look with AI

AI can help you design screens before you write any code.

User Flow

Design Review

Step 6: Build & Test Your Application

Building with AI

This is where AI becomes your coding partner. You'll describe what you want, AI writes the code, and you make sure it works.

Development Checklist

Setup

• ☐Created GitHub repository
• ☐Set up project files
• ☐Can run the app locally

Core Features

• ☐Home screen displays correctly
• ☐Data entry form works
• ☐Data saves to database
• ☐Can view saved data

Testing

• ☐Tested on different browsers
• ☐Tested on phone
• ☐Had a friend try it without help
• ☐Fixed all major bugs

Bug Tracking

Getting Your App Online

Step 7: Collect Data & Observations

Before You Start Collecting

Data Collection Plan

Monitoring Your Data

Observation Log

Step 8: Analyze Data & Draw Conclusions

Analyzing with AI

AI can help you understand what your data means.

Creating Visualizations

Key Findings

Writing Your Conclusion

Reflection Questions

What I Learned:

• About my topic: [Fill in]
• About building apps: [Fill in]
• About working with AI: [Fill in]

Project Completion Checklist

Step 1: Define Project

• ☐Research question finalized
• ☐Hypothesis written
• ☐Variables identified

Step 2: Tools Setup

• ☐All accounts created
• ☐Software installed
• ☐AI assistant ready

Step 3: App Design

• ☐User stories written
• ☐Features listed
• ☐MVP defined

Step 4: Tech Stack

• ☐Technologies chosen
• ☐Understand why each was picked

Step 5: UI/UX

• ☐Screens designed
• ☐User flow mapped

Step 6: Build & Test

• ☐App built and working
• ☐Tested thoroughly
• ☐Deployed online

Step 7: Data Collection

• ☐Data collection complete
• ☐Observations logged
• ☐Data quality verified

Step 8: Analysis

• ☐Data analyzed
• ☐Visualizations created
• ☐Conclusion written

Glossary

Tips for Working with AI

Good Prompting Habits

1. Be specific — "Help me build a form" is okay. "Help me build a form that collects name, age, and 3 multiple choice answers" is better.
2. Give context — Remind AI what you're building and what you've already done.
3. Ask for explanations — Don't just copy code. Ask "explain what this does" so you learn.
4. Break big tasks into small ones — Instead of "build my whole app," go feature by feature.
5. Verify AI's work — AI makes mistakes. Test everything. If something seems wrong, ask AI to double-check.

When AI Gets Something Wrong

Resources

AI Assistants

• Claude — Great for coding help and explanations
• GitHub Copilot — Suggests code as you type (free for students)

Free Hosting

• GitHub Pages — Free hosting for simple websites
• Replit — Build and host in your browser
• Vercel — Free hosting with easy setup

Learning More

• Khan Academy — Free coding courses
• Scratch — Visual coding to learn concepts
• freeCodeCamp — Free full courses