OVERVIEW

Activity Overview:

In this activity students will learn about algorithmic bias through real-world examples. They will then learn the four main principals of ethical AI and apply those to the real-world examples, examining the ethical implications of each example. Students will apply their knowledge as they discuss how bias and ethics in relation to AI can have impacts on our society.

Meta description

• Subject Area: Computer Science, Technology, Engineering, Advisement or other general enrichment classes
• Grade Level : 6-8, 9-12
• Computer Science Domains:
  • Algorithms and Programming
  • Impacts of Computing
• Computer Science Principles:
  • Fostering an Inclusive Computing Culture
  • Communicating About Computing
• Materials:
  • Website
• Considerations:
  • Teachers should be familiar with ways that bias can be present in AI models and have some examples/a good understanding of ethical complications related to artificial learning.

Lesson Plan

Overview

In this activity students will learn about algorithmic bias through real-world examples. They will then learn the four main principals of ethical AI and apply those to the real-world examples, examining the ethical implications of each example. Students will apply their knowledge as they discuss how bias and ethics in relation to AI can have impacts on our society.

ASSESSMENT PRE/POST-TEST

What are ways that AI models can be biased? How can bias in AI lead to ethical issues? Name one way that issues with AI ethics can impact society.

OBJECTIVES

Define the term “algorithmic bias” in the classification context. Understand the effect training data has on the accuracy of a machine learning system. Recognize that humans have agency in curating training datasets. Understand how the composition of training data affects the outcome of a supervised machine learning system. Define the term “ethics” in the context of artificial intelligence/machine learning. Understand how bias relates to ethics in machine learning. Identify how data sets can impact the predictions of machine learning systems allowing them to make unethical predictions. Explain the top four principles of ethical artificial intelligence (fairness, transparency, privacy, human centeredness) and why they are important.

CATCH/HOOK

Have students watch this short clip on AI, humans, and bias (https://youtu.be/AUVcF7ehZ28). Ask students to think of one time they have encountered AI bias in their life and write it down. If they can’t think of a specific example of AI bias, they can write about a time they have experienced bias themselves. Invite a few students to share their experiences.

ACTIVITY INSTRUCTIONS

Lesson Resources: Investigating Bias main slides/script and Exploring AI Bias activity from the DAILy Workshop (https://raise.mit.edu/daily/investigating-bias/#exploreBias), Top four ethical principles for artificial intelligence from AIClub (https://www.corp.aiclub.world/post/teaching-ai-ethics-to-high-school-students).

1. Remind students that last time you met, the group worked on building classification models that sorted a piece of data into one group or the other. You also talked a little bit about how AI models can be biased. Today you are going to take a deeper look at bias in machine learning. Go through the slide deck for Investigating Bias.
2. Next, tell students that the class is going to look at different real-world examples of bias and complete the Exploring AI Bias activity.
3. Tell students that you are going to build on their understanding of bias in machine learning by looking at the concept of ethics. Ask the student to define the word ethical, then build on their decision and give examples of situations where an ethical choice was made versus a choice that was not ethical. How do the real-world examples looked at on Google align with ethics? Discuss the top four ethical principles of artificial intelligence. How did students see those reflected in the Google search results? Where else have they encountered these principles?

Supplements

Any items in this section are the property & under the license of their respective owners.

REVIEW

Review the terms algorithmic bias and ethics. Review how AI models can be biased and how that can render models that are not ethical.

STANDARDS