Kavi, the KAVI owl, waving on a book
HomeAcademyAIWho it's forPricingFAQLabsContactGet Started
AI Literacy · Stage 5

Biased Data

An AI learns from the examples it is shown. If those examples are lopsided, its results come out biased and unfair — bias in, bias out.

✦ AI Literacy📘 Stage 5 · S5⏱ ~9 min🧒 Ages 9–10★ Free to start
I can explain biased data leads to biased results.
Start the free lesson →⬇ Printable PDF
Kavi the owl for AI Literacy
📖

Kavi and the Lopsided Box

Kavi the owl trains a machine to spot a good runner. But her box of examples is lopsided — it is stuffed with photos of tall runners and almost no short ones. When a brilliant short runner races in, the machine shrugs: "Not a good runner." Kavi frowns. "That's not fair — it only ever SAW tall runners, so it decided good means tall!" The data was biased, so the result was biased too. "Bias in," hoots Kavi, "bias out. To be fair, a machine needs fair examples."

🎓

What your child will learn

1Balanced data vs biased dataAn AI learns the pattern it sees most in its data — the examples it is shown. Whether that data is balanced or lopsided decides whether its results are fair. Tap each side!
2Balanced data, or biased data?Here are the example sets an AI could learn from. Put each one in the right box. Balanced data shows many groups and kinds fairly. Biased data over-shows one kind and leaves others out. Tap a box!
3Will the results be fair?Read what each AI learned from, then decide if its results will be fair or unfair. Ask: was the data balanced, or lopsided? Pick, then tap to check. Tap Next test for more.
🔍

A peek inside

A machine learns to spot a good runner, but its examples were ONLY tall runners. Will it judge a short runner fairly?
The machine learns the pattern it saw most. Its data was lopsided — all tall runners — so it learned 'good runner = tall' and barely knows short runners. That lopsided data gives a lopsided result: it rates the short runner low, even if she is fast.

Try this at home: What is biased data?

💬

Questions parents ask

+What is biased data?

Biased data is a lopsided set of examples — it over-shows one kind or group and under-shows or leaves out the others. For example, a set of 'doctor' photos that only shows men is biased, because it leaves women out.

+Why does biased data lead to biased results?

An AI learns the pattern it sees most in its data. If the data is lopsided, the AI learns a lopsided pattern. It then works well for the group it saw a lot of, and poorly or unfairly for the group it barely saw. People call this 'bias in, bias out'.

+Can you give a simple example?

A face-unlock AI trained mostly on one group of faces unlocks well for that group but often fails for others. A voice helper trained on one accent struggles with other accents. The unfair result came straight from the biased data.

+How do you make an AI's results fairer?

Train it on balanced, representative data — examples from many groups, kinds and places, in fair amounts. Balanced data in gives fairer results out.

+What age is this 'Biased Data' lesson for?

It is a Stage 5 lesson for roughly ages 10–11 in KAVI Academy's AI-literacy track. It uses a story, sorting and predicting games, with light reading.

+Is the lesson free?

Yes — this overview and the printable PDF handout are free. The full interactive lesson and the endless practice are part of KAVI Academy.

Ready to play and learn?

Start the free interactive lesson, or print the handout for screen-free play.

Start the free lesson →⬇ Download PDF

Part of AI Literacy · Stage 5 · S5 on KAVI Academy — a secular, whole-child learning platform for ages 5–12.