Summary
An AI model can't know what it was never taught, and that gap becomes bias. This lesson unpacks how training data shapes AI answers and challenges students to rewrite prompts that fight stereotypes instead of reinforcing them.
Bellringer: 1800s AI
- Pose the 1800s-trained-AI scenario and ask students to predict its answers about modern careers
- Collect a few verbal predictions before revealing why outdated data produces outdated advice
Instruction
- Define bias and algorithmic bias using plain-language examples
- Walk through the Amazon resume filter and facial recognition case studies
- Introduce feedback loops and demonstrate the biased vs. unbiased prompt comparison
Discussion: Can AI Ever Be Neutral?
- Debate whether any AI system can be truly neutral
- Extend the conversation to real-world stakes like hiring, admissions, and advertising
Activity: The Binge-Watch AI
- Answer questions from the perspective of an AI model trained only on one TV show.
Wrap-Up & Assessment
- Explain the problem of an AI feedback loop.
- List two ways you can fight bias in an AI model.
| Est. Time | Activity | Description |
|---|---|---|
| 5 min | Bellringer: 1800s AI |
|
| 10 min | Instruction |
|
| 10 min | Discussion: Can AI Ever Be Neutral? |
|
| 15 min | Activity: The Binge-Watch AI |
|
| 5 min | Wrap-Up & Assessment |
|
Lesson Prep
The Binge-Watch AI
Answer questions from the perspective of an AI model trained only on one TV show.
Class Presentation
Includes a 17-slide deck with teacher notes. Present from Banzai, download a PDF, or copy it to Google Slides.
Presentations, Ready to Go
- Built-in PresentationPresent full screen, straight from Banzai
- Copy to Google SlidesMake an editable copy for your class
- Download or PrintSave as a PDF or print
- Teacher NotesSlide-by-slide guidance, printable as one handout
AI Literacy
Synthetic Media
Your eyes can no longer be trusted to spot a fake photo or video. Explore how generative AI blurs the line between real and synthetic media, and learn practical techniques—like reverse image search and metadata checks—to trace where content actually came from.
Bias In, Bias Out
An AI model can't know what it was never taught, and that gap becomes bias. This lesson unpacks how training data shapes AI answers and challenges students to rewrite prompts that fight stereotypes instead of reinforcing them.
Using AI Honestly
Using AI isn't automatically cheating—it depends on whether it supports your thinking or replaces it entirely. This lesson helps students distinguish AI support from AI substitution, explore what only humans can offer, and practice disclosing AI use the right way.
How do Banzai lesson plans work? See the FAQ.