R&D flagship · Grades 3–6

Rex & Dino's Big AI Ideas

Not coding. Not tool training. Human judgment. A comic-first AI cognition series helping children understand how AI learns, why it makes mistakes, and why people still matter.

The complete 16-panel Unit 1 comic plus partial lesson guides — free for educators and parents.

Human Judgment — AI cognition cover
Active R&D

Season 1: AI Mistakes & Human Judgment — SELKID's strategic growth engine, built on 229+ published classroom resources and the Rex & Dino IP.

The framework

Learn → Fail → Judge

01

How AI learns

Children understand that AI systems learn from examples — not magic.

02

Why AI fails

Bad data, missing groups, false confidence, and repeated copying.

03

Why people matter

Checking evidence, context, and taking responsibility.

Season One curriculum

Six units. One judgment framework.

Unit 1 — Training Data Unit 1

Training Data

How AI learns from examples

Rex and Dino discover that a smart trail only understands the examples it has seen.

Buy on TPT →
Unit 2 — Data Quality Unit 2

Data Quality

When examples go wrong

A clever berry basket learns the wrong lesson when examples become rotten, fake, old, or mislabeled.

Buy on TPT →
Unit 3 — AI Bias Unit 3

AI Bias

Unfair patterns in the data

A polite mirror seems smart, but its recommendations become unfair because the examples are unbalanced.

Buy on TPT →
Unit 4 — AI Hallucination Unit 4

AI Hallucination

Confidence is not evidence

Echo Cave answers with confidence, but Rex and Dino learn that sounding sure is not the same as being right.

Buy on TPT →
Unit 5 — Model Collapse Unit 5

Model Collapse

When copies drift from truth

Copied maps slowly lose contact with the real forest, showing how repeated copies can drift from truth.

Buy on TPT →
Unit 6 — Human Judgment Unit 6

Human Judgment

Why people still matter

Atlas Tortoise checks the fastest path against real conditions, proving that useful tools still need people.

Buy on TPT →
Children should not grow up believing AI is magic, neutral, or always right. They should grow up with stories that make AI understandable, questionable, and human-centered.