Parallel Lines, Real Bias: How AI Draws Geometry and the World

Students use AI image generators to explore how parallel lines are shown in real-world scenes. They compare AI images with their own drawings or photos and reflect on what is included, left out, or assumed.

This activity helps students see how AI imagines geometry—and the world. When students ask AI to draw parallel lines in real places, they see more than just math. They notice what the AI includes or leaves out—people, buildings, even who’s visible in the setting. This shows what the AI thinks is “normal.” It also shows how math and culture mix in ways we don’t always expect.

Instructions

  1. Open an AI image generator (like DALL·E, Gemini, or Bing Image Creator).
  2. Type a prompt like:
    “Draw a street with two clearly parallel lines in the sidewalk. Include people walking, signs, buildings, and stores. Make it look like a real place where people live.”
  3. Look closely at the image the AI gives you. Are the lines really parallel? What else do you notice?
  4. Ask follow-up prompts if needed, like:
    “Add more detail to the people” or “Make it look like a place in my community”
  5. Save the AI-generated image.
  6. Now, take a real photo or draw your own version of parallel lines in a street or walkway.
  7. Put the two images side by side: the AI image and your real/drawn one.
  8. Compare them. Look at the lines, the setting, and who or what is shown.

Conscientization

Reading the world through this activity

  • Are the lines in the AI image truly parallel? How do you know?
  • Who is shown in the image? Who is left out?
  • What does the AI seem to think is “normal” in a place?

Praxis

Reflection leading to transformation

  • What could happen if AI keeps showing places this way?
  • How might these images affect how people see real communities?
  • What would you want AI makers to understand from this activity?

Dialogue

Ongoing discussion

  • Share your two images with a partner or group. What differences do you see?
  • What patterns show up across different people’s AI images?
  • How do these patterns connect to how real places and people are often shown—or not shown—in media?
Scroll to Top