Module 9 ā Build Simple AI Projects! š ļø
intermediate30 XP
Your First AI Showcase! š
Use Google Teachable Machine to build and show off a real AI project ā no code needed
Your First AI Showcase! š
You've Learned So Much ā Now Show It Off!
Time to build something impressive to show your friends and family. We'll use Google Teachable Machine to create a real AI project.
šÆ Project Ideas (Pick One!)
Idea 1: The Mood Detector šš¢š
Train AI to recognise 3 facial expressions:
- Happy face ā plays happy music
- Sad face ā sends an encouraging message
- Surprised face ā shows a funny GIF
Steps:
- Go to teachablemachine.withgoogle.com
- Choose "Image Project" ā "Standard image model"
- Create 3 classes: Happy, Sad, Surprised
- Take 50+ photos of each expression (use webcam!)
- Click "Train Model"
- Test it ā does it read YOUR mood?
Idea 2: Rock Paper Scissors AI āāāļø
Train AI to recognise hand gestures:
- Fist = Rock
- Flat hand = Paper
- Two fingers = Scissors
The AI plays against you in real time!
Idea 3: The Homework Helper Detector š
Train AI to detect:
- "Working hard" (sitting at desk with book)
- "Distracted" (phone in hand š )
- "Taking a break" (stretching, standing)
Idea 4: The Sound Detector š
Use the Audio project:
- Train it to recognise: a clap, a snap, "hey computer"
- Use this to trigger different actions!
š Your Project Checklist
ā” Chose a project idea
ā” Created 2-3 categories in Teachable Machine
ā” Collected 50+ examples per category
ā” Trained the model (watched accuracy improve!)
ā” Tested with NEW examples it hasn't seen
ā” Shared with someone and showed them how it works!
ā” BONUS: Export and embed in a website or Scratch project
š Level Up: Connect to Scratch!
Teachable Machine works with Scratch! You can:
- Export your model
- Import it into Scratch using the Teachable Machine Scratch extension
- Make sprites react to your AI predictions
- Build a full interactive game powered by YOUR AI!
š” What Real AI Engineers Do
Congratulations ā you've just done EVERYTHING a real AI engineer does:
- ā Defined the problem (what should AI detect?)
- ā Collected training data (photos/sounds)
- ā Trained the model
- ā Evaluated performance (accuracy)
- ā Deployed it (shared/used it)
- ā Iterated (improved by adding more data)
You are officially an AI builder! š¤šØ
Quick check
01/2
What does Teachable Machine's 'accuracy' percentage tell you?