AI Literacy for Families · Pitjantjatjara

Week 2 — Bias-ku munu Deepfake / AI bias and deepfakes

Pitj: Two problem big. One: AI nintiringu data-ngka iriti — wiya anangu tjuṯa-ngka, mainly whitefella country-ngka, English-ngka. Palunya AI nyangatja palya wangkanyi whitefella worldview. Two: Deepfake — picture, video, voice, AI palyantja palyaringu, anangu maḻa kulini “anangu mulamula” — wiya, AI palyantja.

Eng: Two big problems. First: AI was trained on data mostly from one country, one language, one worldview. So its answers carry that bias. Second: deepfakes — fake pictures, videos, voices made by AI that look real but are not. Both problems need the same defence: check sideways before believing.

Discussion / Wangkanyi mara

  1. AI nintiringu one country-ngka, one language-ngka — nyaa missing? Whose voice missing?
  2. Online — picture nyaa AI palyantja kulira? Nyaa tjukurpa kuḻini?
  3. Deepfake — kuṟu (humour) palya? Kuṟu wiya (scam) palya? Kuṟu wiya (harassment) palya? Same treatment?
  4. Deepfake nyuntu-tjara (or hoa-tjara) online — nyuntu nyaa palya? Nyuntu-tjana watjala first?

Activity / Palyani: “Bias audit + deepfake hunt”

Part A (20 min) — Bias audit. Ask an AI image generator (ChatGPT image, Gemini, free one) to draw images for ten neutral prompts: “a CEO, a nurse, a criminal, a scientist, a person cleaning, a wedding, a beautiful house, a homeless person, a teenager studying, a family at dinner.” No adjectives. Look together: who is shown? Gender, skin, age, body? Who is missing? Anangu missing? Aboriginal people shown — or not?

Part B (20 min) — Deepfake hunt. Go to MIT “Detect Fakes” and work through examples. Then scroll your normal social feed for 10 minutes hunting for AI-generated images. Compare notes — what gave them away?

Parent resources

Reflection

Picture kutju nyuntu nyakuningi week-ngka — kuwari palya nyakunytjaku again. Name one piece of media you saw this week you now want to re-examine.