1.2. Lesson 2: Biases in AI¶
Time Estimate: 45 minutes
1.2.1. Introduction and Goals¶
Can a computer that uses AI be biased? Yes, even though computers are machines, they are not free from the bias of the people who program them and the input data generated by humans. In this lesson, you will learn about AI bias and some ways to reduce or eliminate biases in your programs.
Learning Goals: At the end of this lesson, you will be able to:
- Discuss how technology that use AI have biases.
- Identify wake words, intents, and utterances of existing Alexa skills built in MIT App Inventor.
Language Goals: At the end of this lesson, you will be able to:
- Describe in writing benefits of using Alexa during space travel.
- Explain in writing the importance of identifying and addressing bias in AI.
1.2.2. Learning Activities¶
In this lesson, you will be exploring the concept of biases in AI. This video will provide an anecdote of how biases can be introduced in computer devices. As you watch, write down some issues you have heard about Voice AI (like Amazon Alexa or Apple Siri).
Discussion: Biases in Computer Science
Computing innovations can have the same biases as people because of the way the algorithms are written. This can be a problem when the computer is used for things like medicine, business, and science. The programmers need to try to reduce the bias in the algorithms to make sure that computing innovations do not discriminate against groups of people. In this video, you will see more examples of algorithmic bias. As you watch, think about some ways that you could combat biases.
Activity: Good Morning, Space!
In this activity, you will explore the App Inventor interface by using and modifying a pre-built Alexa skill. Before going to the interface, review the steps of the Alexa dialogue.
This is an example of a dialogue with Alexa. Sarah wants Alexa to turn on the lights.
The wake word is a word that causes the device to begin recording a user's request. In this example, Sarah uses Alexa as the wake word.
The utterance is the command or question a user will speak to trigger a specific action as part of the skill. Sarah uses turn up the lights to indicate that she wants the lights to turn on.
The intent is the desired response. Sarah wanted Alexa to turn on the lights - this is Sarah’s intent. Along with turning on the lights, Alexa responded saying the intent.
Now that you have reviewed wake word, utterance, and intent, open Alexa’s App Inventor and log in with your Google account.
Next, you will need to open the Good Morning, Space skill. Follow along with your teacher to explore the interface.
Check for Understand: Good Morning Space
After identifying the wake word, utterances, and intent of the Good Morning, Space skill, answer the following questions.
Q-3: What was the wake word used in the Good Morning, Space skill?
Q-5: What is the intent of the Good Morning, Space skill?
Activity: Modifying Good Morning, Space!
In the previous activity, you explored the components of the Alexa skill, Good Morning, Space! Now that you have been introduced to the App Inventor interface, try modifying the program. Can you change the utterance? How about the response from Alexa? Don’t be afraid of breaking the skill - you can always get a fresh copy.
Q-6: What did you modify about the skill?
Q-7: Did you encounter any challenges? If so, what did you do?
In the next lesson, you will help build another Alexa skill and learn more about the Artemis space program. explore how AI is currently being used in space and some of the ways it could be used in the future. For now, reflect on what you have learned in this lesson.
Q-8: Why is it important to recognize and address bias in AI?
Q-9: As you supply more information to Alexa, what types of biases might you unintentionally “teach” it?
Q-10: List 3 possible benefits of using an Alexa while traveling in space.