AI UNIT 1 • STAGE 3 OF 5

Where Answers Come From

Training data, in plain language, and why it shapes what the AI can say about your people

← All AI Units ← Stage 2 AI Unit 1 Stage 4 →
Unit
STEP 1

Back to the elder's question

In Stage 2 you saw the AI do well on some questions and poorly on others, especially about your own nation. This stage answers why.

It comes down to one idea: training data. An AI learns by being shown an enormous pile of writing, mostly text collected from the internet, and finding patterns in it. It doesn't look anything up. It repeats and remixes the patterns it saw.

STEP 2

Ask it where its knowledge came from

Put the question to the AI directly.

Try these

Where did your information about Native peoples come from? Did Native people write most of what you learned about us? Why do you know more about some topics than others?

Notice whether it admits the limits of what it was trained on.

STEP 3

Whose writing is in the pile?

Here's the part that matters for your community. The AI is only as good as what people wrote down and put online. So ask yourself:

  • If lots of accurate writing exists online about a topic, the AI usually does well.
  • If little exists, or what exists is wrong or stereotyped, the AI repeats those gaps and mistakes, confidently.
  • Much of what's online about Native peoples was written about us, by outsiders, not by us.

💡 Key idea

An AI can't know your nation better than the internet does. When your community's own voices aren't in the training data, the AI fills the gap with guesses and stereotypes.

STEP 4

Connect it in your notes

Look back at your Stage 2 examples. In your Field Notes, explain training data in your own words, and connect it to what you saw: where were the AI's answers thin, and whose voices were missing from the "pile"?

🪶 This is data sovereignty, starting

Who gets to write the record about your people? Who owns that data? You're now asking the exact questions the later AI units are built around.

Ask the AI
🛡️ Guided classroom sandbox. Your messages are saved for your teacher.
Ask the AI where its knowledge comes from, and whose voices shaped it.
Field Notes

Explain what you're learning in your own words:

  • 1. In your own words, what is "training data"?
  • 2. Who do you think wrote most of what the AI learned about Native peoples?
  • 3. Looking at Stage 2, where were the AI's answers thin, and whose voices were missing?

Saved automatically.