---
title: Understanding Bias
units: [Problem Solving With AI]
summary: AI can be biased.
weight: 60
---

{{% param summary %}}

## Today's Objectives
- Analyze AI-generated responses to identify bias and assess its impact on fairness.
- Experiment with AI prompts and data inputs to explore how they affect AI bias.
- Explain how AI bias emerges from training data and influences AI-generated outputs.
- Identify strategies for mitigating bias in AI systems and evaluating ethical considerations.

## Lesson Overview
### Why can AI be biased?
{{< collapse summary="Click here to reveal the answer." >}}

AI can be biased because it learns from patterns in data, but that data isn't always balanced or fair. If certain perspectives or groups are overrepresented, AI will reflect those patterns in its responses. This can lead to AI making assumptions that don’t accurately represent the full picture.

{{% define "Bias" %}}

{{</ collapse >}}

### How can you reduce AI bias?
{{< collapse summary="Click here to reveal the answer." >}}

You can:

- Improve training data to be more representative.
- Use prompting strategies to mitigate bias.

{{% define "Alignment" %}}

{{</ collapse >}}

## Assignment
{{% instructions-unit-journal-update %}}
{{% unit-journal-define-terms "Bias" "Alignment" %}}

### AI Audit
As you complete Code.org, record the following.

#### Level 1
- Original Prompt
- Observations
    - AI Response
    - Bias Identified
    - Why This Bias Might Exist
    - How to Reduce This Bias
- Refined Prompt

#### Levels 3 & 4
- Role (Hospital Triage Nurse, College Admissions Officer, or Software Engineering Hiring Manager)
- Your Selections
    - Which two people you selected
    - What qualities or information you prioritized
- AI's Selections
    - Which two people AI selected and whether it matched what you chose
    - What qualities or information AI prioritized

#### Level 6
- Whether AI's selections changed when you told it which qualities to prioritize and why or why not

### Reflection
After completing Code.org, record the following:

- Examples of where you saw AI bias
- What you can do to help mitigate it

{{% instructions-code-org-update %}}
{{% unit-journal-question-of-the-day question="What is AI bias and how does it emerge?" %}}