1

2

Before Lesson 1:

Take the Pre Survey

Intro to AI Pre Survey

tinyurl.com/aiEDUintroToAiPreCourse

3

4

Lesson 1

Defining the undefined

Introducing: My AI Workbook

​

What is it?

• One stop shop for lesson overviews, important vocabulary, and all of the links you will need
• A place for structuring your final project

​

​

Make your own copy:

tinyurl.com/aiStudentWB

​

6

→ Explore two of the applications on this page:

• Final project examples

​

For each app you choose, think about the following:

1. What is the topic?
2. What kinds of questions does it have and how do you input your answers?
3. Aside from the survey, what else do you see on the page?

​

​

​

7

Student Workbook: Page 6

🔥 Warmup

Not everyone agrees on the definition of AI, but one good definition is “a man-made system that exercises problem-solving or decision-making abilities normally associated with humans”.

​

8

🎯 The Point

Quick, Draw! Live - Rules

Goal: Guess your classmate’s drawing in 45 seconds or less.

​

Rules for the person drawing:

1. When it is your turn, come to the front of the room and choose a card randomly from the pile.
2. Read the item on the card, show it to your teacher, and take 5 seconds to think about how you will draw it.
3. You have 45 seconds to draw whatever is on the card.
4. As you draw, the rest of the class will shout out what they think you are drawing.

9

Quick, Draw! Live - Rules

Goal: Guess your classmate’s drawing in 45 seconds or less.

​

Rules for the rest of the class:

1. While someone is drawing, if you think you know what they are drawing shout it out.
2. You don’t have to wait until the drawing is finished. You can start guessing as soon as the timer starts.
3. If you guess correctly, you get a point.

10

Quick, Draw! Live - Let’s play!

1. Choose a card and read it.
2. Take 5 seconds to think about how you will draw it.
3. When the timer starts, you have 45 seconds to draw.
4. If you hear the correct answer from one of your classmates, stop drawing.

11

Explore AI Examples: Quick, Draw!

→ Explore the AI below:

• Quick, Draw!

​

While you play, think about these questions:

1. How does Quick, Draw! work (i.e. how does it know what you are drawing)?
2. Compared to humans, was the application better or worse at guessing your drawings?

12

No consensus

Identify the definition you think is best and explain why.

→ There is no single agreed upon definition of AI.

​

→ Check out the list of AI definitions and consider:

1. What words, phrases, or ideas do these definitions seem to have in common?
2. How are they different from each other?
3. Which definition do you feel is best? Why?

13

Student Workbook: Page 85

If you finish early…

An ⭐ AI is a man-made system, such as a computer program, that exercises problem-solving or decision-making abilities normally associated with humans.

14

Return to projects

→ Explore the Intro to AI - Clubs project at this link:

• Final project examples

​

1. How is this application using AI?

15

Student Workbook: Page 6

What is one example of AI that you see or use in your everyday life and how does it use AI?

16

AI in your everyday life

​

Apps like Netflix, Amazon, and Spotify are called recommendation systems.

​

Recommendation systems use AI to make recommendations to users.

​

What kinds of things do Netflix and Amazon recommend?

17

Final Projects

Check out the Clubs app again:

​

1. How is this application using AI?

18

It is a recommendation system. It recommends what club a user should join.

Final Projects

Final project goal: build a recommendation system about a topic that you are passionate about.

19

Reflection

1. Now that you have seen some final project examples, what types of topics might you use for your project?

​

​

20

Lesson 2

Can’t spell AI without data

→ You’re in charge of loaning money for a large bank. You make money by charging interest on the loan, but only if the person pays it back over time.

1. What information would you like to know about a person in order to decide if they should get a loan?

​

​

22

Student Workbook: Page 7

🔥 Warmup

AI finds patterns in data and works best with lots and lots of high quality data.

23

🎯 The Point

🚨 Vocab alert!

• ⭐ Data is any form of recorded information.
• Examples include:

24

17

A fingerprint

A photo of a face

An age

Data is the lifeblood of AI.

(You can’t have AI without data.)

25

AI finds patterns in data

26

Data

Decisions

Example: AI finds patterns in data

27

Data

​

what TikTok videos you watch & how long you watch them

Decisions

​

videos on your TikTok For You page

What do you think Snapchat uses this data for?

28

Data Snapchat Uses

​

• date of birth
• type of phone
• most and least used filters
• your interests
• how often you message each friend
• how often you open Snapchat

Decisions

​

???

Garbage in, garbage out

29

Wrong or Unimportant Data

Bad or Biased

Decisions

Loan Prediction Game

→ Apps like Rupee Factory use AI to approve or deny a loan for up to a couple thousand bucks in just 10 seconds. Let’s see how they work.

​

Directions:

• Approve or deny a $1000 loan for each person
• Be ready to justify your response
• You’ll see who did or didn’t pay back their loan at the end!

30

Person 1: Aishwarya

• never keeps her phone charged
• uses an iPhone 5c
• took 1 minute to type in her birth date
• orders takeout once per week
• gets a new phone every year
• made 4 typos when filling out her application

31

Would you loan Aishwarya $1000???

​

Why or why not?

Person 2: Nicky

• never keeps their phone charged
• uses a Google Pixel 5a
• took 4 seconds to type in their birth date
• never orders takeout
• has not bought a new phone in 4 years
• made 0 typos when filling out their application

32

Would you loan Nicky $1000???

​

Why or why not?

Person 3: Sham Sham

• charges her phone every night
• uses an iPhone 13 Pro Max
• took 11 seconds to type in her birth date
• orders take out 4 - 5 times per week
• buys a new phone 2 times per year
• made 12 typos when filling out her application

33

Would you loan Sham Sham $1000???

​

Why or why not?

Loan results

After 1 month:

• Aishwarya: paid back $800
• Nicky: paid back $1000
• Sham Sham: paid back $200

​

​

​

​

What could explain these results?

34

Explore Loan App Data Collection

→ Investigate each data point:

🔗Instant Loan App Data Points

​

For each data point:

1. Why does the loan app collect it?
2. How do you think they collect it?
3. Would you be comfortable with a bank using this data to decide if you can get a loan?

35

Loan apps like these collect 20,000+ data points. We only saw 10 of them. What other data do you think they collect?

36

Explain why you think these loan apps are or are not fair.

37

Reflection

→ Instant loan apps are commonly used in India and China.

1. How do you think these instant loan apps affect the average user’s life?

​

​

38

Lesson 3

Songza ‘bout right

→ Songza was a music recommendation service that operated from 2007 - 2016. Unlike Spotify or Apple Music, Songza relied on human “music experts” to create playlists for individual users.

1. What information would you give to a “music expert” from Songza in order for them to make a good playlist for you?

40

Student Workbook: Page 8

🔥 Warmup

Just like humans, AI recommendation systems can show bias in their recommendations.

41

🎯 The Point

Write Playlists

• Create a playlist of at least 15 songs for your user persona

​

Then, answer:

1. From 1 - 10, how difficult was it to create this playlist?
2. From 1 - 10, how similar are you to the user persona you got?
3. Do you feel like you would be able to make the same quality of playlist for everyone?

42

Compare & contrast playlists

• Form a group with everyone else who had the same user persona
• Share & compare your playlists

​

Then, answer:

1. Why did you choose the songs you did?
2. What kind of assumptions did you make?
3. What would have made it easier to make this playlist?

43

Revisit playlists - More Information

Extra information

• User persona 1 likes country music
• User persona 2 likes rap and hip hop
• User persona 3 likes opera
• User persona 4 likes classical music

​

​

44

Revisit playlists

Discuss:

1. What is your favorite genre of music? Tell your group about it.
2. What, if anything, did you change about your playlist after seeing the new information? Why?
3. Based on what you know about the user’s taste and each group member’s taste, who in the group would be the best person to make this playlist?

​

​

45

Reflection

1. Would AI be better or worse at making a playlist? Why?
2. Is it possible for AI to be biased?

46

Lesson 4

Two Wrongs Make a Right

→ Imagine one of your friends comes to you for advice. They are trying to buy a gift for someone you don’t know and they want recommendations from you.

​

1. What questions would you ask your friend?
2. How might it be different if the gift is for one of your close friends?

​

​

48

Student Workbook: Page 9

🔥 Warmup

AI can be both biased and inaccurate, but even imperfect AI systems can be useful.

49

🎯 The Point

Recommendation Systems - Amazon

​

Amazon uses an AI-powered recommendation system to recommend products to users.

​

How does it know what you would like to buy?

50

Recommendation Systems - Amazon

​

• Powered by your user data.
• What you’ve purchased before
• How you’ve rated other purchases
• Even what items you have looked at even if you didn’t buy them

​

51

Recommendation Systems - Amazon

​

• Powered by other user’s data
• Which other users have the same likes and dislikes at you?
• What else have they purchased?
• What is trending in the moment?

​

Amazon’s recommendation system isn’t perfect!

52

When things go wrong

Open the Amazon Mistakes page in your workbook.

• For each question, note which item you think Amazon recommended

​

Think about the following questions:

1. Why did Amazon make these weird recommendations?
2. Have you ever seen a weird recommendation while using any AI recommendation system?

53

Student Workbook: Page 92

Another Example

Recommendation system mistakes can have serious consequences.

​

YouTube:

• Similar to Amazon but videos
• Recommends what videos users will like based on their viewing history.
• Knows what is trending

54

YouTube Action Plan

You are an executive at YouTube and you have received news that there is a problem with the site.

​

The recommendation system is directing a lot of users to conspiracy theory videos.

​

It is your job to create an action plan to fix this.

55

YouTube Action Plan

Plan should include:

1. What do you think caused the problem?
2. What are some possible ramifications?
3. Who is responsible for fixing it?
4. How would you fix the problem?

56

Reflection

1. What role does the government play in the YouTube problem?

57

Lesson 5

Get inspired!

​

Open one of the sample programs again and try to guess what type of person created that project. Fill in the 3 sentence stems below with your answer.

​

• ❝ The _ app was developed by a _ ❞
• ❝ The developer is interested in _ ❞
• ❝ They made the app because _ ❞

59

Student Workbook: Page 10

🔥 Warmup

AI can help solve problems you are passionate about.

60

🎯 The Point

Project Steps

61

⭐ We are here

Project Goals

• Build an AI-powered recommendation system.
• Around a topic you are interested in.
• Be creative and share your passions with others.
• Real-world application that can be used to help other people even after the course ends.

​

62

Intro to AI Clubs App

• What is this project meant to help with?
• What data does this program use?
• How is the data collected?

​

63

Rapid Brainstorming

Think about…

1. Are there any movies, books, shows, or video games that you love?
2. What do you do in your free time?
3. Is there a cause that you care about?
4. Have you learned anything in school that was really interesting?
5. Are there places you have traveled to or want to travel to?

64

Rapid Brainstorming

Write down as many topics and ideas as you can think of.

​

1. What do you do in your free time?
2. Are there any movies, books, shows, or video games that you love?
3. Is there a cause that you care about?
4. Have you learned anything in school that was really interesting?
5. Are there places you have traveled to or want to travel to?

65

Concept Mapping

• A concept map is a visual representation of information
• Used to illustrate connections between different concepts that relate to one specific topic

66

Example

Concept Map

67

Introducing: the Toolkit

• Try these techniques if you’re not sure how to get started
• Feel free to skip it if you’re comfortable with the task

68

Concept maps

Try the second concept map with your second favorite topic

Make 2 concept maps. For each map:

• Write one of your topics at the center of each concept map
• Explore that topic by mapping it out
• Use the tools in the toolkit if you need a jumpstart

69

If you finish early…

Share your maps

Think about the following questions while look at your partner’s maps:

1. What are some questions you have about each of your partner’s topics?
2. Which ideas on the maps do you find the most intriguing?

​

70

Reflection

1. You made concept maps for two topics. Does either topic stand out to you right now?

71

Lesson 6

Solving your problem with AI? Priceless.

Go back to the sample program again. After you’ve finished it, answer the following questions.

1. How accurate was the program’s recommendation? Be sure to look at the accuracy number that appears after you’ve filled out the survey.
2. Noting the program’s accuracy, do you think this recommendation system is useful?

73

Student Workbook: Page 11

🔥 Warmup

Successful AI applications begin with narrow, well-thought-out problem statements.

74

🎯 The Point

Example Project

Let’s check it out: tinyurl.com/IntroToAi-clubs

75

Example Project

1. What might be limiting the accuracy of this application?
2. How could you make it better?

76

Our Constraints

• A ⭐ constraint is a restriction that limits an idea.
• All projects have constraints.
• Our constraints:
• Building a recommendation system
• Using survey data to build our project and train our AI application

​

77

Real World Constraints

• In a previous activity, you acted as a music recommendation system. What were the constraints that limited your ability to make a playlist?
• Think about one of the applications we have talked about so far:
• Amazon
• YouTube
• Loan app

What are some constraints you can identify in these programs?

​

​

78

Create problem statements

Start thinking about the “who” of your problem - who does this help, hurt, or affect neutrally?

Develop a problem statement for at least two of the topics you identified.

​

A ⭐ problem statement is a brief description of an issue that a project aims to address.

​

A good problem statement is specific and realistic.

79

If you finish early…

Turn and Talk

Share your problem statements with your neighbor

1. What is the specific problem being solved here?
2. Why is this an important problem to solve?
3. Is this a good use of AI?

80

81

Lesson 7

My AI ate my homework

→ Look at these faces and try to describe…

1. What emotion is being expressed?
2. How can you tell?

83

A

B

Student Workbook: Page 12

🔥 Warmup

→ Here’s some extra data about the person

1. How does having this extra data change your answer?

84

🔥 Warmup

Limited or poor quality data can lead to biased AI: AI that works differently for different groups of people.

85

🎯 The Point

How do humans identify emotions?

Maybe you noticed the shape of someone’s mouth

​

​

​

​

86

How do humans identify emotions?

Or the angle of their eyebrows

​

​

​

​

87

How do humans identify emotions?

These are facial features

​

​

​

​

88

🚨 Vocab alert!

• ⭐ A feature is a measurable piece of data that can be used by an AI to make a decision
• Example: features of a video game include genre and price

89

Genre sci fi western

Price cheap expensive

AI features aren’t the same as human ones.

• An AI doesn’t even know what a face is!

​

• It would have to look at thousands or millions of images of faces before it can learn to detect different emotions.

90

Let’s identify some features!

→ For each topic, explain which features you would like to use to make your decision.

91

Buying a car

92

What features would you like to know about a car to decide if you want to buy it?

Buying a car

93

What would happen if the salesperson lied about the features of the car and you didn’t realize it?

Choosing a best friend

94

What features would you like to know about a person to decide if they could be your best friend?

Choosing a best friend

95

What would happen if the person was acting fake and you didn’t realize it?

Remember, data is the lifeblood of AI.

(You can’t have AI without data.)

96

Predict test scores

→ In a moment, you will get into groups and receive student test score information cards.

​

✄ When you receive your sheet of cards, cut them out.

97

Predict test scores

Each card has a few different data points that tell you something about the student.

​

98

Predict test scores

​

1. Sort each card based on what grade you think the student will receive on the final exam (A, B, C, D, or F).

​

1. For each datapoint you have, explain how you think it would affect a student’s test score.

​

1. What data do you wish you had to help improve your prediction?

99

Imagine these were the scores…

100

How accurate were your predictions?

Imagine these were the scores…

101

Choose one prediction you got wrong. Why do you think you got it wrong?

Test scores

→ What if you had all of these data points? Would your prediction be more or less accurate? Why?

102

Test scores

103

→ Grades were based on:

​

• Last quiz score

​

• Hours studied

​

• Teacher quality

AI is used in many ways in education…

104

1. What features do you think the AI used to detect cheating?
2. How do you feel about colleges using this AI?

Reflect

1. How would you improve the AI that detects cheating during an exam? Use the word feature in your answer.

105

• Expand our understanding of AI and bias
• Investigate AI’s impacts on employment

106

👉🏿 Up next!

Lesson 8

What starts in bias ends in bias

1. What might explain the difference in how well an AI application can recognize a “lighter male” compared with a “darker female”? Use the word data in your answer.

108

Student Workbook: Page 13

🔥 Warmup

Here’s the full graph, if you’re curious…

109

AI can absorb biases from the real world, and companies and governments must take responsibility for those biases.

110

🎯 The Point

Investigate hiring bias

→ Choose one of the three examples of hiring bias:

1. Amazon’s Gender Biased Hiring AI
2. LinkedIn’s Biased Job Matching AI
3. Uber’s Racially Biased Face Recognition System

​

111

Investigate hiring bias

→ For your article, answer the questions below:

1. How did the AI affect who each company hired or fired?
2. What do you notice about the race breakdown of the company?
3. What do you notice about the gender breakdown of the company?
4. How do you think the biased AI and the breakdown of the team could be connected?

112

Create Your Position Poster

Who is responsible for correcting bias in AI?

​

Use the template to create a position poster based on the biased AI you investigated.

​

• Describe the bias you are trying to address.
• Explain the composition of the team that created the application
• Draw a conclusion about who is most responsible for fixing the bias

113

Reflect

1. How did the job-related AI you looked at mirror a real world bias?

114

Lesson 9

#SquadGoals

→ Answer with your team:

1. What makes a successful team in business, sports, gaming, or another area?

​

​

116

Student Workbook: Page 14

🔥 Warmup

Project Steps

117

⭐ We are here

Teams that divide work based on each person’s strengths build the best solutions.

118

🎯 The Point

Find Common Ground

Try brainstorming some new problems that everyone in your squad might be interested in.

→ Share your problem statements from the previous lesson with your teammates.

​

→ In your workbook:

1. Decide on one problem statement you would like to work on as a team.

119

If you finish early…

Example Squad Agreement

120

Write Your Team Agreement

Come up with a team name together!

→ Rotate to each of the 3 stations to determine your roles.

​

→ At each station:

• Add the name of each of your squadmates to the table
• Assign a role to each squadmate according to that station’s rules
• Explain why your squad decided on that role for that squadmate

​

​

121

3 rounds of 8 minutes

If you finish early…

Share out

→ Let me know if you have each role:

• Opener
• Tech Runner
• Closer
• Presenter
• Social Media Aficionado
• Email Expert
• Charmer
• Texter
• Googler

And finally… point to your backup!

122

Lesson 10

Who am I designing for?

→ Choose one of the two web apps below to explore:

• What anime should you watch next?
• Which self care strategy might work for you?

​

Then, answer the questions below:

1. What is each web app’s topic question?
2. What are they recommending and what are the different options you can get?
3. Who is the audience of this app?

124

Student Workbook: Page 15

🔥 Warmup

Thinking about the users of an AI application helps you understand what data to collect in order to solve the problem.

125

🎯 The Point

Our Constraints

• A ⭐ constraint is a restriction that limits an idea.
• All projects have constraints.
• Our constraints:
• Building a recommendation system
• Using survey data to build our project and train our AI application

​

126

Problems we can’t solve within our constraints

• Recommending someone a pair of shoes based on an image of the outfit they are wearing right now
• Why? We can only collect survey data for our recommendation system
• How to adapt this problem: have people answer questions about what type of clothes they are wearing instead

127

Survey

1. What style of pants are you wearing?
2. Are you wearing a watch?

Problems we can’t solve within our constraints

• Figuring out how to solve climate change
• Why? This does not involve recommending something for a specific person.
• How do adapt this problem: recommend the best option for a person to reduce their carbon footprint, such as flying less

​

128

🚨 Vocab alert!

• A ⭐ user persona is a fictional person who would want to use an AI application, usually used as a tool to imagine the needs of real users.
• There is no one right way to create a user persona. Let’s explore three different approaches!

129

Example User Persona

130

Notes:

• Includes Josh’s income and occupation
• Explains what Josh enjoys doing
• Includes personality traits

Create Your User Persona

Brainstorm a user persona you are definitely NOT designing for.

→ On your chart paper, draw your user persona and add notes to explain what you’re drawing. Make sure to add:

• Name
• Identities
• Personality traits
• Goals
• Fears
• What they do during a typical day
• Anything else you can think of!

131

If you finish early…

User Persona Gallery Walk

→ Hang up your user persona poster, then examine other squads’ user personas, noting down:

1. Which user personas would also use your app?

​

​

132

Reflect

1. What data would you like to collect about your users to make the best recommendations for them?

133

Lesson 11

Oh yeah, it’s all coming together

Many social media platforms and news sites like Facebook, Reddit, BuzzFeed, and Instagram include quizzes.

1. Why do you think quizzes are so popular on social media?

135

Student Workbook: Page 16

🔥 Warmup

Your recommendation system needs to collect data about relevant features of a person in order to provide meaningful results for users.

136

🎯 The Point

🚨 Vocab reminder!

• A ⭐ feature is a measurable piece of data that can be used by an AI to make a decision.
• Examples: a person’s favorite dessert, the number of legs an animal has, and the colors in an image.

137

BuzzFeed Quiz Exploration

Suggest two more questions you could ask to improve the accuracy of the quiz.

→ Choose one of the three quizzes:

• Page 1: Where will you travel next?
• Page 11: Which anime hero are you?
• Page 23: Which zodiac sign are you?

​

→ After examining the quiz, answer:

1. What is the quiz’s target question?
2. What feature does each question try to measure about a person?
3. How accurate do you think this quiz would be for you and why?

138

If you finish early…

🚨 Constraint Alert!

• Reminder: a ⭐ constraint is a restriction that limits an idea.
• When you build your Google Form in Module 5, you will only be able to use TWO types of questions

Multiple Choice Dropdown

​

139

For each question…

• List all the possible answers to your questions
• ✅ Which snack do you prefer? [list options]
• ❌ What is your life story?
• Limit the number of answers to each question to 10 or fewer.
• ✅ Which of these 7 moods are you most often in? [list options]
• ❌ What is your favorite number?

140

What to Ask

Add some more questions that are out of the box - like the BuzzFeed quizzes!

→ Return to your user persona. For each ⭐ feature you said you wanted to measure, add the following to your workbook:

1. What question would you need to ask someone in order to collect that piece of data?
2. What are the possible answers to that question?

​

🚨🚨 Try to think of at least 8 questions.

141

If you finish early…

Reflect

1. How accurate do you think your recommendation system will be for the average user and why?

142

143

Lesson 12

Metaphoring is believing

Think about a skill you learned in the past, like a musical instrument, a sport, or a task at work or school.

1. What actions did you take in order to learn this skill?

145

Student Workbook: Page 17

🔥 Warmup

Machines and humans learn skills in a similar way: by training repeatedly and getting better over time.

146

🎯 The Point

🚨 Vocab alert!

• ⭐ Machine learning is a type of AI that recognizes patterns in data and draws conclusions based on those patterns.

147

The Four Components of Machine Learning

• ⭐ Data: any form of recorded information
• ⭐ Algorithm: a set of rules used to solve a problem.
• ⭐ Model: a representation of a real-world decision-making or problem-solving process.
• ⭐ Prediction: an estimate based on prior data.

148

A Metaphor for Machine Learning

→ This is Rohith. We’ll use Rohith’s journey to becoming a world-class baseball pitcher as a metaphor for machine learning.

​

→ While Rohith is not a machine, we can break down how and what he learns to understand the four components of machine learning: data, algorithm, model, and prediction.

149

Data

150

→ Spin

→ Speed

→ Pitch

→ Left or right hander

→ Batting record

→ Hit or strike?

Algorithm

151

Throw 10,000 fastballs against hundreds of batters

Throw 10,000 curveballs against hundreds of batters

Throw 10,000 changeups against hundreds of batters

Model

152

Muscle Memory

Flexibility

Knowledge of which pitches to use when

Prediction

153

A left-hander with a 3.80 average and two strikes so far… I think a fastball at 90 mph straight down the center will strike him out. Let’s try it.

Write your own metaphor!

Brainstorm multiple predictions for your metaphor

→ You just saw one metaphor for machine learning. In a pair or on your own:

• Develop your own metaphor for machine learning
• Draw out and explain each of the four components on your handout: data, algorithm, model, and prediction

​

→ You can use the skill you thought of in the warm-up or come up with a different example.

154

If you finish early…

Let’s share our metaphors!

155

Reflect

1. What do machine learning and human learning have in common?

156

• Develop a plan to minimize bias in your AI application
• Start planning to ⭐ pitch your idea to your classmates

157

👉🏿 Up next!

Lesson 13

Kryptonite for bias

→ You’re an engineer at Amazon overseeing the launch of a new AI application. The application uses people’s resumes to recommend who to hire. During the launch, you discover that only 15% of the candidates recommended by the application are women.

1. How would you feel in this situation?
2. What would you do in this situation? (e.g. what questions would you ask, what would you do with the application)

​

159

Student Workbook: Page 18

🔥 Warmup

Responsible AI designers build a concrete plan for minimizing bias in their AI application before they create it.

160

🎯 The Point

Project Steps

161

⭐ We are here

Building an Anti-Bias Plan

→ Check out at least 2 of the biased AI applications below.

• COMPAS
• Facebook ads
• Apple Card

​

Then, consider:

1. What went wrong?
2. What is ONE thing my squad can do better to avoid this bias?

162

🚨 Vocab alert!

• A ⭐ pitch is a brief presentation used to persuade others to support an idea

163

Video Focus Question

1. What makes this video an example of a pitch?

164

Which pitch should we watch?

165

Hot Seat

Scrub Daddy

Animal Fitness Tracker

Start Pitch Deck

→ One person in your squad should make a copy of the pitch deck template, then share it with everyone in your squad.

​

→ Once you have your copy:

• Complete slides 1 - 4 together

166

Share out

1. What is one anti-bias strategy you added to your pitch?

167

• Researching others who have tried to solve the same problem as you
• Continuing to add to your pitch deck

168

👉🏿 Up next!

Lesson 14

Part of something greater

→ Examine the site you received.

1. Would this site be a trustworthy source for your research topic “AI that recommends anime”?
2. Why or why not?

170

Student Workbook: Page 19

🔥 Warmup

Project Steps

171

⭐ We are here

In order to be widely used, a new AI recommendation system needs to address a new problem or address an old problem in a new way.

172

🎯 The Point

Research Focus Questions

1. How have others tried to tackle this problem before us?
2. How is our recommendation system unique, special, new, or creative?

173

Landscape Research

→ With your team, conduct research into:

1. How have people tried to tackle this problem before us?
2. How is our recommendation system unique or creative?

​

Try these sentence stems for Googling:

• ❝ solutions for _ ❞
• ❝ ai and _ ❞
• ❝ companies solving _ ❞
• ❝ recommendations for _ ❞
• ❝ problems with _ ❞

174

Update Pitch Deck

→ Locate your team’s copy of the pitch deck from the previous lesson.

​

→ Once you have your copy:

• Complete slides 5 - 7 together

175

Reflect

1. Circle the top 5 most important keywords from your research.
2. Summarize your research in a sentence or two using those keywords.

176

• Finalize & personalize your pitch deck
• Practice your pitch

177

👉🏿 Up next!

Lesson 15

Practice makes permanent

→ Locate your team’s pitch deck.

1. What sections do you need to complete in the pitch deck?
2. What improvements would you like to make to it?

179

Student Workbook: Page 20

🔥 Warmup

Successful product pitches feel effortless and interesting because they have been practiced.

180

🎯 The Point

Finish building your pitch deck

Write cue cards, change the graphics of your pitch, or plan a short skit to open your pitch!

→ Complete any remaining slides in your pitch deck.

181

If you finish early…

Practice pitching

Work on your cue cards to help you remember what to say

→ Practice your pitch out loud, paying special attention to:

• starting the presentation
• ending it
• switching speakers

182

If you finish early…

Wait, who’s doing what?

Indicate if your role during a presentation is:

• tech runner
• opener
• closer
• presenter

183

Reflect

1. What do you hope your pitch will communicate to your audience?

184

185

Lesson 16

One puzzle piece said to the other…

→ Imagine a company called Viewclear AI is developing a facial recognition software that it plans to sell to the government of your city. The city will use the AI application to track every person in the top 10 busiest areas of the city.

1. What questions could you ask to figure out if this use of AI was overall beneficial or harmful?

187

Student Workbook: Page 21

🔥 Warmup

All AI applications have risks and benefits.

188

🎯 The Point

AI Critique Jigsaw

→ In a moment, you will get into groups to explore an AI application you choose.

​

→ On your card, your group name is at the top.

189

AI Critique Jigsaw

→ In a moment, you will get into groups to explore an AI application you choose.

​

→ Your assigned Question number is at the bottom.

190

AI Critique Jigsaw

→ In your group, you will create a Public Service Announcement (PSA) to help people make a decision about whether the tech you choose is overall beneficial or harmful.

​

→ But first, let’s check out the tech…

191

AI Critique Jigsaw

→ In your groups (e.g. Hot Sauce), choose 1 of the 4 AI applications. Examine:

• Benefits of this tech
• Risks of this tech

​

→ Based on your question number, answer one additional question:

1. What makes this technology AI?
2. What data is used to train the AI application?
3. What decisions or predictions does the AI make?
4. Who is this AI designed for? Who uses it?

192

AI Critique Jigsaw

→ Now, regroup based on your question number (e.g. everyone who had Question #4 work together).

​

Share your answers and discuss:

• Areas of common ground
• Areas of disagreement

193

AI Critique Jigsaw

→ Return to your original groups (e.g. Hot Sauce).

​

Start working on your PSA!

• Use your answers to Questions 1 - 4 to help you explain the tech
• Outline its benefits and risks
• Don’t try to convince the reader of your opinion – just give them the facts

194

Reflect

1. How impressed were you with the technology you read about? Did it seem futuristic, obvious, unexpected?

195

• Finish your PSA posters
• Look over your classmates’ PSAs and share your opinions

196

👉🏿 Up next!

Lesson 17

… that’s a fresh ‘fit

→ Check out your PSA poster from the previous lesson.

1. What has your group completed on your PSA poster?
2. What still needs to be done?

198

Student Workbook: Page 22

🔥 Warmup

People critically evaluate AI applications according to their risks and benefits based on their unique identities and perspectives.

199

🎯 The Point

Complete your PSA

Look up some other info sources about your technology to add to your PSA.

→ Once you are back in your tech groups from the previous lesson, complete your PSA poster.

​

→ In about 20 minutes, you will hang up your poster so others can see your thoughts.

200

If you finish early…

PSA Gallery Walk

→ Place a sticky note on the side of their poster that matches your thoughts:

201

👎🏽 Left side: more harmful overall

👍🏾 Right side: more beneficial overall

PSA Gallery Walk

→ Walk around the room and view the other groups’ posters.

​

→ Place a sticky note on the side of their poster that matches your thoughts:

• 👍🏾 Right side: more beneficial overall
• 👎🏽 Left side: more harmful overall

202

Which PSAs had the most positive and negative reactions? Why?

203

Reflect

1. What questions can you ask to figure out whether an AI application is overall beneficial or harmful?

204

Lesson 18

Pitch-a-palooza!

Project Steps

206

⭐ We are here

Student Workbook: Page 23

An effective pitch for an AI application convinces your audience that the benefits of your technology outweigh the risks.

207

🎯 The Point

🚨🚨 Your feedback matters 🚨🚨

→ After each pitch, record your thoughts on 2 separate sticky notes:

208

One benefit of the recommendation system

One risk of the recommendation system

🚨🚨 Your feedback matters 🚨🚨

​

→ Structure your thoughts using these sentence stems:

209

Reflect after you pitch

→ After your squad pitches, reflect:

1. How do you feel your pitch went out of 10?
2. What do you feel went well?
3. What is one thing you would change if you could try again?

210

3-2-1 pitch!

211

Double check: did your squad collect all their feedback sticky notes?

212

Reflection

→ Consider each of the pitches you saw today.

1. Which one resonated with you the most or was the most interesting to you and why?

213

• Revise your project based on the feedback you received from other squads

214

👉🏿 Up next!

Lesson 19

Mirror, mirror on the wall

→ Think back to your squad’s pitch from the last lesson. You can check out your reflection for a refresher.

1. What is one piece of advice you would give to someone about to pitch their own project?

216

Student Workbook: Page 24

🔥 Warmup

Practicing iterative design helps refine AI applications over time.

217

🎯 The Point

Improving Your Designs with Iteration

Video Focus Question

Why is the process of iterative design useful for someone developing a product?

​

218

Review your feedback

→ Use the feedback sorting chart to organize the feedback you received

​

• Good stuff: what your squad did well with the project
• Improvement: what could be made better or expanded
• Garbage can: feedback that just isn’t useful

219

Review your feedback

→ Sort your feedback using the chart.

→ Then, on your chart:

1. Star the top 3 items in Good Stuff
2. Star the top 3 items in Suggestions

​

→ Then:

1. Summarize the strengths of your project using the top 3 items you starred in Good Stuff
2. Summarize how your project could be improved using the top 3 items you starred in Suggestions

​

220

If at first you don’t succeed…

→ Iterate, iterate again! Choose a focus area and explain your plan for improving your project.

221

Iterate!

→ Choose one area of your AI application you would like to improve:

• Problem statement
• Data

​

→ Jot down how you will change that area based on what your peers said.

​

​

​

​

222

Reflect

1. How did you practice iterative design today?

223

• Figure out the difference between muffins & cupcakes
• Explore the complexities of identifying and collecting data

224

👉🏿 Up next!

225

Lesson 20

Muster for data

→ Imagine you are building an AI application that detects the difference between muffins and cupcakes.

​

1. What types of data could it use to make decisions?
2. How would you collect that data?

227

Student Workbook: Page 25

🔥 Warmup

There are many different types of data, and appropriate data collection methods lead to better AI applications.

228

🎯 The Point

Types of Data

229

Biometric Data

Data that comes from measuring a person’s physical or behavioral characteristics

​

Types of Data

230

Numerical Data

Data that is made up of numbers

​

Types of Data

231

Visual Data

Data that is made up of things like images or videos

​

Types of Data

232

Audio Data

Data that consists of sounds

​

Types of Data

233

Textual Data

Written or printed words, sentences, and other strings of characters

​

Data Collection Plans

→ Imagine you are an employee at tech nonprofit UAIDE. You are asked to build a custom application for a local business.

​

→ The first step in building a new app is creating a Data Collection Plan (DCP).

​

→ A DCP outlines what kind of data a new app needs and how to collect it.

234

Data Collection Plans

Create a data collection plan for your assigned app. A good DCP answers these questions:

1. What will you name the app?
2. At least two types of data you could use to create the app?
3. For each type of data you identified, how would you collect that data?
4. What will be difficult about collecting that data?
5. A description of how the data might be biased and how you plan to avoid that.

235

Improved Data Collection Plans

Take turns sharing your data collection plan with your new group

• Give each other feedback
• Using what you like from each DCP, edit your DCP accordingly

236

DCP gallery walk

→ In a moment, everyone will get a chance to try every pair’s plans.

​

→ As you work, you will leave a note that answers this question:

• After reading this plan, are their any questions you would still have?

237

Reflect

1. Which part of your plan took the longest to complete? Why do you think that is?

238

Lesson 21

With great power…

→ Unsafe & Sound Security, an AI-based technology company, is testing their new indoor camera system. They are offering $20/week to anyone who agrees to place cameras in every room in their home. They will provide a privacy policy that guarantees the videos and images they collect will never be viewed by a human, only by AI.

​

1. Would you participate and why or why not?
2. Does it matter that it won’t be watched by humans?

240

Student Workbook: Page 26

🔥 Warmup

Many companies collect huge amounts of data about people with minimal consent in order to power their AIs.

241

🎯 The Point

A day in the life

→ You are researching the amount of private data collected on members of a small community called Bison, NY.

​

→ You will receive a report on one community member’s average day.

​

242

A day in the life

→ Read through your community member’s story.

​

→ Any time you think data is being collected, write down the type of data on a sticky note.

​

​

​

243

A day in the life

→ Add all of your sticky notes to the data collection wall

​

​

​

244

1. If the companies from the stories share and/or sell data to each other, what would they know about each person at the end of the day?

​

1. What would they know about the community of Bison, NY as a whole?

245

Who is responsible?

→ Multiple people and institutions involved in data privacy

• Tech companies who collect the data
• Government agencies who could have the power to regulate data collection
• Individual users who allow their data to be shared.

246

1. Who is more responsible for ensuring consumers are aware of how their data is used - tech companies or users?

​

1. What else that could be done to prevent data like this from being shared?

247

Reflect

Imagine you are the community member from your story.

1. How would you change your behavior after examining the data collected about you, if at all?

248

Lesson 22

Sharing isn’t always caring

1. Which of these pieces of information would you be willing to tell a stranger?
2. Which would you definitely NOT be willing to tell them?

250

Student Workbook: Page 27

🔥 Warmup

Project Steps

251

⭐ We are here

When AI applications require personal user data, the responsibility of data privacy is shared between users, companies, and government.

252

🎯 The Point

Explore AI Examples

→ Explore one of the applications on this page:

• Final project examples

​

For the app you choose, think about the following:

1. Do you think the questions in this project are asking for sensitive data?
2. Are there any situations where a user might be uncomfortable sharing the data?

253

Student Workbook: Page 27

We value your privacy … no, seriously, we do

→ Before you collect your data, it’s important to get in the mindset of your user personas and think about the sensitivity of the data you want to collect.

254

How can you figure out how sensitive certain data is?

255

Consider…

Is this something people would normally share with:

• strangers?
• close friends?
• family?
• no one?

256

Consider…

What would happen if someone found out this data was associated with a certain person?

257

Consider…

Is the data:

• medical?
• financial?

258

Consider…

Does the data reveal someone’s demographic information, such as:

• name?
• gender?
• race?
• national origin?
• political opinions?
• religious beliefs?
• sexual orientation?

259

Finalize your data collection plan

→ For each piece of data you want to collect for your project:

• Rate how sensitive it is (1 - 10)
• Explain your reasoning

​

→ Then add your conclusions:

1. What is the average rating of how sensitive this data is?
2. How sensitive is the data you are collecting overall?
3. Are there any changes you would like to make to your data collection plan?

260

Example privacy disclaimer

Thank you for taking the time to share some of your thoughts and help others. This survey asks you about some of your experiences in order to help others find a good self care strategy that fits who they are. The survey does ask some personal questions, such as “what is a difficult emotion you have struggled with in the past?” We understand this information is sensitive and we will only use this data to make good recommendations for others about how to overcome that type of difficult emotion.

We appreciate your time in taking this survey. Your bravery allows us to create positive dialogue around the issue of mental health!

​

261

Write privacy disclaimer

→ In your workbook, write a short note that you will include in your survey:

• thank the person for sharing their data
• acknowledge how sensitive the data being shared is
• explain how the data will be used

262

Create an action plan

1. What are your roles during data collection?
2. How many people do you think you will be able to get to take the survey?

263

Lesson 23

Now let’s get in formation

→ Explore and compete at least 2 of the surveys below:

• 🔗 Chinchilla survey
• 🔗 Pet survey
• 🔗 Career Survey

​

Then, consider:

1. Can you answer in a way the survey creator wasn’t expecting?
2. Did any of the questions make you concerned about data privacy?
3. Can you submit the survey while it’s still incomplete?

​

​

265

Student Workbook: Page 28

🔥 Warmup

High quality data collection tools, like well-designed surveys, lead to more accurate AI applications.

266

🎯 The Point

🚨🚨 Form Constraints Alert 🚨🚨

• You MUST make every question required
• You can only use two question types:
• dropdown
• multiple choice

​

​

If you do not follow these guidelines, your recommendation system literally will not work.

267

Form Creation Demo

→ Let’s walk through the steps together.

​

One person from your team

1. Creates a new form
2. Adds your teammates as collaborators

268

Create Your Google Form

Add images or gifs to your form, change the theme, or share with a neighboring squad

→ One person on your squad should:

• Make a copy of the template form
• Add every squadmate as a collaborator

​

→ Then, everyone should work on:

• Copying the disclaimer from The Fine Print section
• Copying your questions from the What to Ask section
• As the LAST question, adding your target question

269

If you finish early…

Generating a Share Link

→ Each Google Form has TWO links:

• One 🔨 edit link for adding questions and formatting
• One ✉️ share link for taking the survey

​

270

🔨 Edit Link

271

✉️ Share Link

272

Form Share Link Demo

→ Let’s walk through the steps together:

1. Click the Send button
2. Select the Link tab and check Shorten URL
3. Copy the link
4. (optional) Use 🔗 tinyurl.com to create a shorter link

273

Reflect

1. How do you think people will respond to your survey as you collect data?

274

275

Lesson 24

AI Olympics

What do you think would be easiest to teach an AI-powered robot and why?

1. How to take a selfie
2. How to solve a calculus problem
3. How to pack a box

​

​

277

Student Workbook: Page 29

🔥 Warmup

Successful AI is built around an algorithm with clearly sequenced steps and properly collected data.

278

🎯 The Point

The AI Olympics

→ It’s time to test the strength of your AI understanding in the AI Olympics!

​

→ There will be 3 team events where your team could win:

1st place = Gold Medal

2nd place = Silver Medal

3rd place = Bronze Medal

279

AI Olympics

Event 1: The Selfie Algorithm

​

​

Event 2: AI Biathlon

​

​

Event 3: Accuracy Archery

280

Event 1: The Selfie

Algorithm

Goal: Be the first to write a detailed, step-by-step selfie algorithm

​

1. Assemble with your team near a part of the whiteboard.
2. As a team, write down a selfie algorithm - the steps someone would follow to take a selfie.
3. When you think you have a working algorithm, send one athlete to the front of the room.
4. That athlete must read their algorithm to a volunteer from the other team who will follow the instructions and see if they work.
1. Hint: Volunteers can take each step very literally

281

Medal Ceremony

282

Reflect

1. Do you think a long, more detailed algorithm is always better than a short, less detailed algorithm? Why or why not?

283

Lesson 25

AI Olympics: 2 fast 2 curious

→ Two competing companies, BookSlug and Heads Down Books, provide book recommendations to users. The companies have very different approaches to making recommendations:

• BookSlug spends 80% of their time collecting excellent data and 20% creating their algorithm.
• Heads Down spends 20% of their time collecting data and 80% of their time perfecting their algorithm.

​

1. In your opinion, which company will make better recommendations? Why?

​

​

285

Student Workbook: Page 30

🔥 Warmup

AI can learn to identify human behavior by repeatedly studying how humans act.

286

🎯 The Point

AI Olympics

Event 1: The Selfie Algorithm

​

​

Event 2: AI Biathlon

​

​

Event 3: Accuracy Archery

287

AI Biathlon pt. 1

Data Sprint

Goal: Illustrate the process of training an AI model

​

1. Every group will get a copy of the Data Sprint chart.
2. On the data sprint chart handout, fill in the chart displayed on the next slide with the words from the list provided.
3. When you think you have a correct chart, raise your hand for a check.
4. If your chart is correct, you can tag in Group 2 from your team. They will receive a handout with the next set of instructions.

288

The data sprint chart shown here illustrates the process of training and testing a model using data.

​

You will fill in the chart with the correct terms which are displayed on the following slide.

Data Sprint

289

Data Sprint

290

Correctly fill in the boxes with these terms:

• Clean training data
• Model
• Prediction
• Collect test data
• Algorithm
• Clean test data
• Collect training data

Collect training data

AI Biathlon pt. 2

Beat the Machine

Background:

A lab at MIT trained an AI to predict human behavior from watching TV characters’ body language.

​

The machine was trained on 1000s of episodes of different TV show. It was then asked to predict whether a character will hug, high five, yell, or shake hands.

​

The machine was right 43% of the time.

291

AI Biathlon pt. 2

Beat the Machine

Goal: Predict human behavior better than MIT’s AI

​

1. Each team has 100 points
2. For each round, you will see a still image from a TV show.
3. With your teammates, bet some or all of your points on the whether the character(s) will hug, high five, yell, or shake hands.
4. If your guess is correct, you will win the same amount of points that you bet and they will be added to your total.
5. If you beat the machine (get at least ⅗ correct), your team gets an extra 10 points.

292

AI Biathlon pt. 2

Beat the Machine

1. HUG
2. HIVE FIVE
3. SHAKE HANDS
4. YELL

293

Beat the Machine

294

AI Biathlon pt. 2

Beat the Machine

1. HUG
2. HIVE FIVE
3. SHAKE HANDS
4. SLAP

295

Beat the Machine

296

AI Biathlon pt. 2

Beat the Machine

1. HUG
2. HIVE FIVE
3. SHAKE HANDS
4. YELL

297

Beat the Machine

298

AI Biathlon pt. 2

Beat the Machine

1. HUG
2. HIVE FIVE
3. SHAKE HANDS
4. YELL

299

Beat the Machine

300

AI Biathlon pt. 2

Beat the Machine

1. HUG
2. HIVE FIVE
3. SHAKE HANDS
4. YELL

301

Beat the Machine

302

Medal Ceremony

303

Reflect

Think about today’s games:

1. In general, do you think humans are better or worse at predicting human behavior?
2. Why is it difficult for AI to predict what humans will do?

304

Lesson 26

AI Olympics: end game

→ In Beat the Machine, you tried to make more accurate predictions than an AI application.

1. What information would have made it easier for you to guess correctly?
2. What would the AI have to do to get more accurate?

​

​

306

Student Workbook: Page 31

🔥 Warmup

Like humans, AI isn’t perfectly accurate - but imperfect AI can still be beneficial.

307

🎯 The Point

AI Olympics

Event 1: The Selfie Algorithm

​

​

Event 2: AI Biathlon

​

​

Event 3: Accuracy Archery

308

Accuracy Archery

Goal: Successfully build and accurately throw a paper airplane

​

1. Divide your team into 2 groups.

​

1. Assemble Group 1 from each team around a table with several blank pieces of paper. Each team’s Group 2 is not involved yet.

​

1. You have 8 minutes to create a paper airplane. You can test several versions before you decide on a final model.

309

Accuracy Archery

Goal: Successfully build and accurately throw a paper airplane

​

1. When time is up, each member of Group 2 will take a turn trying to land their airplane on the target. The first archer should begin 10 paces back from the target. After that, every team member can move 1 pace closer to the target before they throw.

​

1. Teams will receive one point for every time they land on the target. The lowest score is 0/5 and the highest is 5/5.

310

Reflect

1. What was difficult about that event?
2. How would your accuracy improve if we played more rounds?
3. How does this relate to the accuracy of your projects?

311

Final

Medal Ceremony

312

Reflect

• Next, you will use the data you have collected to create an AI model.
• You will examine your model’s accuracy at that point
• Accuracy might be from perfect, just like your paper airplane throwing skills.
• An imperfect application still gives you a useful prediction.

​

313

Lesson 27

Your new favorite emoji

→ Check out the states dataset. Look for missing values, wrong values, and repeated values.

1. How many issues do you notice? What are they?

​

​

​

Examples of errors:

• Indiana’s abbreviation is ZZZZZ instead of IN
• Mississippi is missing its name
• nebraska.gov is copied into multiple rows

315

Student Workbook: Page 32

🔥 Warmup

Project Steps

316

⭐ We are here

Clean, high quality data powers useful public tools like AI-based web applications.

317

🎯 The Point

Data Cleaning

→ Now it’s time to clean the data you’ve collected for your project

​

1. Are there any missing values?
• If there are only a few missing values and you have a lot of data, delete the row containing the missing value.
2. Is there a column for quiz score?
• If so, it needs to be deleted.
3. Are there any question types not allowed, such as short answer?
• If so, these columns need to be deleted.

318

Hugging Face app

→ We are using a platform called Hugging Face to build and share our web apps

​

→ Create a Hugging Face web app by following the step-by-step instructions in your workbook.

319

Status Check

1. Where are you at with your app? Is it up and running?

320

• Continue working on your web app!

321

👉🏿 Up next!

Lesson 28

Judging an app by its cover

→ Look at this Hugging Face web app and answer the following questions:

1. What do you think is the target question of this web app?
2. What is one important takeaway you learned from exploring this web app?

​

323

Student Workbook: Page 33

🔥 Warmup

Useful public tools like AI-based web applications are well named and clearly explained.

324

🎯 The Point

Hugging Face app continued

​

Continue working on your web app. By the end of this time, your team should:

1. Get your web app up and running.
2. Write info.md to explain your web app.

​

If you finish early, jot down:

• How could you extend your project to use robotics?
• How could you extend your project to use image recognition?

325

Reflect

As you get prepared to present your hard work, consider:

1. How is your app? Are there any issues? Is it up and running?
2. What is your group role when presenting?

326

• Get ready to present your web app to your classmates

327

👉🏿 Up next!

Lesson 29

Practice makes permanent: the remix

1. Write 10 words that would help someone understand everything they need to know about your web app.

​

329

Student Workbook: Page 34

🔥 Warmup

The goal of presenting a public tool like an AI-based web application is to inspire users to trust it.

​

330

🎯 The Point

Cue cards

Cue cards will help you guide your presentation. Use the cue card template to create cards that you think will help you with your presentation.

• You don’t have to use the ideas on the first page (but you can)
• Blank cue cards on page 2

331

Practice, practice, practice

​

Write a short skit to open your presentation or add a section to info.md

Practice the presentation you designed in your cue cards out loud.

332

If you finish early…

Reflect

1. What do you hope your presentation will communicate about your project?

333

• Think about how AI-powered applications impact different stakeholders.
• Predict how user reactions to an app might influence iteration.

334

👉🏿 Up next!

335

Lesson 30

I object!

→ In this lesson, you will examine the journey of an AI application called Speech2Face, which was first released in 2019. The application uses AI to generate images of faces based on a recording of speech.

1. What is one way you think this application could be beneficial to society?
2. How might it be harmful?

337

Student Workbook: Page 35

🔥 Warmup

People respond differently to AI applications based on their unique identities and perspectives.

338

🎯 The Point

🚨 Vocab alert!

• A ⭐ stakeholder is a person with an interest in a specific AI application, such as a government or a user of the application

339

Speech2Face Virtual Timeline

340

The hype

The backlash

The response

Speech2Face Narratives

→ In this activity, you will examine narratives written by different stakeholders in the Speech2Face application.

341

Speech2Face Virtual Timeline

Consider who you agree with the most as well as who you are the most similar to.

→ Read the narratives from each stakeholder. As you read:

• Note areas of agreement and disagreement among stakeholders

​

→ After reading The Response, consider:

1. What changes did the Speech2Face team make in response to the backlash?
2. How well do you think the team balanced the concerns of the four stakeholders involved?

342

If you finish early…

Hashtags and Headlines

→ Choose 1 of the stakeholders from either The Hype or The Backlash. Get in character as that stakeholder!

​

→ As that stakeholder, explain how you feel about The Response from the Speech2Face team in:

• 3 headlines
• 10 hashtags

343

Reflect

1. Why do you think the Speech2Face application isn’t commonly used by the public?

344

• Present your AI application ‼️
• Evaluate each AI application as a pickling business owner, a variety streamer, a nun, and the list goes on

345

👉🏿 Up next!

Lesson 31

This calls for a demonstration

→ In this lesson, you will adopt the perspective of a user persona. As the user persona you got:

1. Brainstorm 3 things you love.
2. Brainstorm 3 things you hate.

347

Student Workbook: Page 36

🔥 Warmup

Who here is …

• under 30?
• over 30?
• male?
• female?
• unspecified gender?
• employed?
• a student?
• a business owner?

348

Project Steps

349

⭐ We are here

Advocating for the accuracy and ethical responsibility of your AI application allows users to critically evaluate it.

350

🎯 The Point

For the rest of this lesson, “I” and “me” means you as your user persona!

351

App presentations!

→ As you watch each presentation, keep track of applications you (as your user persona) would be likely or unlikely to use.

352

For apps you would use:

• ❝ I would likely use _ because _ ❞
• ❝ I think I would get excited about _ because _ ❞
• ❝ The _ application would be great for me because _ ❞

For apps you would NOT use:

• ❝ _ isn’t my style because _ ❞
• ❝ I would probably never use the _ application because _ ❞
• ❝ The _ application wouldn’t be a good fit for me because _ ❞

Present!

353

Reflection

1. Which AI application did you think would be most meaningful to your user persona?

354

• Literally a party
• Try out everyone’s app and leave them a review with your feedback

355

👉🏿 Up next!

Lesson 32

No rest for the predicted

1. What is the purpose of reviews in the app store?
2. Have you ever left a review? If so, what did you say?

357

Student Workbook: Page 37

🔥 Warmup

Providing constructive feedback on an AI application helps its creators improve it through iterative design.

358

🎯 The Point

App reviews gallery walk

→ In a moment, everyone will get a chance to try every squad’s app.

​

→ As you work, you will leave reviews similar to the reviews in the app store.

359

🚨🚨 Your feedback matters 🚨🚨

​

360

Example review

🚨🚨 Your feedback matters 🚨🚨

​

→ Use the sentence stems in your workbook to help structure your thoughts.

361

Take a moment to post the link to your squad’s app on a sheet of paper somewhere around the room.

362

App reviews gallery walk

Add a funny (but not cruel) joke review OR add a review sponsored by a company

→ Walk around and examine each app. Try it out, this time as yourself.

​

→ Once you have tried it, leave a review!

1. Fill out the app store review card
2. Turn your card upside down and leave it in a pile near the app

​

→ Remember that your review will be used by that squad to improve their app in the next lesson.

363

If you finish early…

Summarize

1. How did you feel when reviewing apps? Powerful? Unsure? Critical? Nervous? Why?
2. Complete the statement: “Most apps in this room are _.”

364

• Thinking about what makes a review useful
• Iterating one last time

365

👉🏿 Up next!

Lesson 33

So, we meet again…

→ Check out one of the Yelp reviews, then answer the question below.

1. What valuable information does this review contain for the owner of the restaurant or site?

367

Student Workbook: Page 38

🔥 Warmup

The functionality and ethical standard of an AI application should be improved over time using iterative design.

368

🎯 The Point

Sorting Reviews

→ It’s a good idea to sort reviews so that you can pull out themes in the feedback you received.

​

→ You have some options for sorting your reviews:

• Reviews we mostly agree with vs reviews we mostly disagree with
• Good stuff and suggestions for improvement
• By star rating
• Something else!

369

Label Your Review Categories

→ One example might be:

370

Reviews we mostly agree with

Reviews that don’t align with our view

Garbage can (reviews that just aren’t useful)

Sort Your Reviews

Identify things that no one mentioned in their reviews – what does this tell you about your app?

→ Use the categories you identified to read through and sort your reviews.

​

→ Once you’re done sorting, consider:

1. Which review(s) do you agree most with? Are there any you can throw out?
2. What were the strengths of your app?
3. What was most commonly requested for your app?

371

If you finish early…

🚨 Vocab reminder!

• ⭐ Iterative design means repeatedly prototyping, testing, and refining a product with the goal of improving it.

372

Iterate!

Consider how your app could expand to use robots or image recognition.

→ Choose one of the three focus areas to review and improve based on the feedback you received:

• Problem statement: more specific, expanded, consider something else?
• Data collection plan: low accuracy, didn’t work for some people?
• Web app: improved explanation, more context, working better?

373

If you finish early…

Reflect

1. Describe one successful part of your app and one area you want to improve on.

374

• Take a look at the bigger picture: beyond your project & into the future of work

375

👉🏿 Up next!

376

Lesson 34

Hiring soothsayers, will pay overtime

→ Certain jobs are more likely than others to be automated (replaced by AI applications) in the future. A recent report found that:

• agricultural equipment operator (e.g. tractor driver) is highly likely to be automated
• mortician is extremely unlikely to be automated

​

1. What might explain this difference?

​

378

Student Workbook: Page 39

🔥 Warmup

Jobs that are more likely to be automated by AI tend to be more repetitive and less social.

379

🎯 The Point

What comes to mind when you think of “automation”?

380

🚨 Vocab alert!

• ⭐ Automation is the process of building and using technologies that operate with minimal human intervention, especially AI-powered technologies.

381

Examples of automation

382

Let’s play…

🤖 Bot or ⛔ not!

​

Rules:

→ I give you 2 different jobs

→ You tell me which one is more likely to be automated within the next 20 years

383

physicist

384

VS

model

385

VS

elementary school teacher

boat mechanic

386

VS

speech-language

pathologist

statistician

387

VS

tax preparer

police officer

388

VS

waitress

building inspector

Explore job descriptions

→ You will examine some real job descriptions, looking for:

• Tasks that are easy to automate using AI
• Tasks that are hard to automate

389

Create help wanted ad

Add a few of your own job responsibilities to your help wanted ad

→ Read through one of the provided job descriptions, then:

• Sort the responsibilities based on what would be easy or hard to automate
• Create a help wanted ad for a human or an AI using the responsibilities
• Use the provided template

390

If you finish early…

Reflect

1. What are 3 core features of jobs that are difficult to automate?

391

• Consider what hiring managers might be looking for in the future of work
• Create your own portfolio to show off your skills & projects

392

👉🏿 Up next!

Lesson 35

One weird trick

→ Choose one of the following scenarios below, then answer the question below.

• You are a Technical Program Manager at Google hiring a new software engineer
• You are the Lead Designer at Nike hiring a new graphic designer
• You are the CEO of your own business hiring a new personal assistant
• You are the Head of Finance at TikTok hiring a new accountant

​

1. Which scenario did you choose?
2. What are you looking for in a candidate when hiring for this job?

​

​

394

Student Workbook: Page 40

🔥 Warmup

Project Steps

395

⭐ We are here

A portfolio helps tell a person’s professional story.

396

🎯 The Point

3 Soft Skills for Success

Video Focus Question

What personal qualities can help you land a job?

397

What is a portfolio? What is it for?

398

🚨 Vocab alert!

• A ⭐ portfolio is a collection of a person’s work, skills, education, and values used to tell their professional story.

399

Example: Simon Sotelo / Weird Wonderful

• Explains more about who Simon is
• Describes the kinds of services offered at Weird Wonderful
• Shows off past work

​

400

🔗 link to Weird Wonderful

Portfolio Explanation

→ Explore a few of the portfolios.

• 🔗 May Li Khoe – Designer and Researcher
• 🔗 Simon Sotelo – Visual Designer
• 🔗 Randi Williams – Roboticist

​

→ Keep track of elements of the portfolios that you like and dislike.

401

Now it’s your turn!

Sketch in pictures or animations for your portfolio!

→ Develop your portfolio using the template:

• Homepage
• Pictures
• Your recommendation system
• What’s important to you
• About Me Page
• Education
• Personal qualities
• Work experience
• Hobbies

​

​

402

If you finish early…

Reflect

→ Think back to the portfolios you explored today.

1. What would you like to add to or change about your portfolio if you had more time?

403

• Believe it or not, the last lesson of this course!
• 🔮 A look into the future to examine possible careers that will be resistant to automation

404

👉🏿 Up next!

Lesson 36

The future is what we make it

Consider the five careers below, then answer the questions:

• 👟 Professional athlete
• 🦷 Dentist
• 🃏 Poker dealer
• 💸 Accountant
• 🥔 Potato farmer

​

1. Which of these careers do you think is most likely to be gone by the year 2040? Why?
2. Which one is least likely to be gone? Why?

​

​

406

Student Workbook: Page 41

🔥 Warmup

An understanding of AI can help guide a career.

407

🎯 The Point

What are some qualities (think soft skills) that might be difficult to automate?

408

Getting started

→ Think of at least 3 different jobs you are curious about.

​

→ For each job, make a prediction about each one’s risk of being automated by AI: low, medium, or high.

409

Explore the future of work

Try to find a job with an automation risk of <1% or >98%.

→ Use the site below to explore each job you thought of.

​

• 🔗 willrobotstakemyjob.com

​

→ As you work, jot down your discoveries about each job.

410

If you finish early…

How likely is it that this job will be automated by the year 2040?

411

extremely unlikely to be automated by 2040

extremely likely to be automated by 2040

⬆️ Add your jobs here! ⬆️

Intro to AI Post Survey

tinyurl.com/aiEDUintroToAiPostCourse

​

​

412

• Transfer the skills you have learned here to other parts of your life
• Learn more about artificial intelligence
• Spread the word about the effects of artificial intelligence on our society

413

👉🏿 Up next!

414