AI Explorers Grades 6 to 8
Build a paper AI, test it, audit it for bias, and learn how generative AI works and where it goes wrong, with Nova as your guide.
Free, start to finish. No card. Every week opens without signing in.
- Grade
- Grades 6 to 8
- Length
- 12 weeks
- Pace
- 6 short lessons a week, about 20 minutes each
- Price
- Free, every week
About this course
AI Explorers Grades 6 to 8 is a 12-week AI literacy course for middle schoolers, guided by Nova, the crew's AI companion drone. Students build Sky Sorter, a paper classifier for telescope snapshots: they encode pictures as numbers, design features, write a decision tree as a flowchart and pseudocode, train it, test it on held-back cards and measure its accuracy. Then they name the three kinds of AI bias, learn how generative AI predicts the next word, why it can be confident and wrong, how to spot synthetic media, and how to weigh privacy trade-offs. The course ends with an AI review board that judges Sky Sorter against the qualities of trustworthy AI. Short daily lessons with checked answers, an unplugged Explorer Lab each week and a Saturday Explorer Quest for the family. No student ever uses a real AI tool in the course. 100% free, start to finish.
Kids will
- Define an AI system as a machine-based system that makes predictions, recommendations or decisions
- Encode pictures and observations as numbers and features
- Write a model as a flowchart and pseudocode, then train, test and refine it
- Name the three kinds of AI bias and find one in a dataset
- Explain how generative AI predicts and why it can be confident and wrong
- Spot synthetic media, weigh privacy trade-offs and judge whether an AI is trustworthy
The 12 weeks
-
Week 1
What Counts as AI?
Objectives, outputs and autonomy
-
Week 2
Sensors and Perception
From raw signals to meaning
-
Week 3
Encoding the World
Pixels, color values and feature tables
-
Week 4
Representation and Reasoning
From feature tables to decision trees
-
Week 5
Rules or Learning?
Hand-written rules versus patterns found in data
-
Week 6
Train, Test, Refine
Measuring how well Sky Sorter really works
-
Week 7
Three Kinds of Bias
When a fair-looking score hides a gap
-
Week 8
How Generative AI Makes Things
From sorting to making
-
Week 9
Confident and Wrong
Why smooth answers still need checking
-
Week 10
Real or Synthetic?
Slow down before you share
-
Week 11
Privacy and Trade-offs
What to share, what to protect
-
Week 12
The AI Review Board
Review, refine and present
More in AI Explorers
Try Week 1 today.
Open the course, pick Day 1 and go. Make a free parent account when you want progress saved for each kid.