1 of 38

1

MAKE BEATS. LEARN CODE. PROMOTE EQUITY.

v1.0

2 of 38

2

BROUGHT TO YOU BY

WITH SUPPORT FROM

3 of 38

3

WHAT IS AI ?

4 of 38

Learn to recognize AI and how AI works so you can use AI effectively in a safe and responsible manner

4

OBJECTIVE

5 of 38

5

WHAT IS AI?

Term	Definition
ARTIFICIAL INTELLIGENCE (AI)	�When a computer mimics human intelligence or behaviour�

6 of 38

6

CAN YOU FIND EXAMPLES OF AI?

Can you think of a time when an app or website

Recommended something to you
Decided what you would see without asking

You should totally check out this new movie

Instructor Dialogue

“Can you think of a time when an app or website

Recommended something to you?
Decided what you would see without asking?”

Instructor Cue

Ask the students to think of examples for each scenario.

Here are examples of each you can share

AI is often used to make recommendations or suggestions

When your favorite video streaming service lists “Recommended movies” that’s done with AI e.g. “Your Next Watch” or “Because you just watched” suggestions on Netflix
When your favorite music streaming service makes a playlist for you, that’s done with AI e.g. “Smart Shuffle” on Spotify
When an online store lists “others also bought” or “products often bought together”

Sometimes it’s more subtle.

Many of the ads that pop up when you are visiting websites are selected based on AI
What appears in your social media feed is often determined by AI

Related articles

https://www.aboutamazon.com/news/retail/amazon-generative-ai-product-search-results-and-descriptions

NOTE: The cartoon image of two people talking was generated by DeepImg on April 10, 2025, (deepimg.ai/ai-image-generator/)

7 of 38

7

CAN YOU FIND EXAMPLES OF AI?

Can you think of an example where a computer, robot, or device acted like a human by

Listening
Talking
Seeing

Moving
Thinking
Creating

Instructor Dialogue

“Can you think of an example where a computer, robot, or device acted like a human by

Listening
Talking
Seeing
Moving
Thinking
Creating

Listening, talking, looking, moving, and creating are all activities performed by humans when you see technology performing these activities it’s usually AI”

Instructor Cue

Ask the students to think of examples for each scenario. This can be done as a class discussion, or they have space in their Student Workbook Activity 1.1 Identify examples where you use AI to write down examples on their own and them you can ask students to share some of the examples they wrote down with the class.

Here are examples for each

Listening

Have you ever asked your phone what the weather is today?
Have you ever asked a device to play a song?

When a computer listens and understands what you say it is using a type of AI called Speech to Text

�Talking

Have you ever been in a car while an app tells you when to turn right or left?
Have you ever had a phone read you your messages?

When a computer reads something out loud, it is using a type of AI called Text to Speech

Seeing

Have you ever used FaceID on a phone? Learning to recognize different faces is done using AI
Have you ever played with an app that changes how you look e.g. puts a hat on your head? changes your hair? Turns you into a cartoon character? That relies on AI to identify the different parts of your body so it can change your appearance

Moving

Have you ever seen a robot that moves like a human - walking, jumping, running? That requires a lot of AI to figure out how to bend joints and shift weight so the robot can balance and not fall over. There are some remarkable humanoid and dog robot examples from Boston Dynamics https://www.youtube.com/@BostonDynamics
In Manufacturing if a robot just does exactly the same thing over and over never changing that doesn’t usually require AI. BUT! If a robot is doing a task and needs to reach to it’s surroundings that is usually done with AI. For example Amazon uses robots to pick up and move shelves so people packing boxes don’t have to walk around finding objects, those robots use AI to make sure they don’t crash into each other and to make sure they don’t bump into a person (Amazon’s first autonomous robot)

Thinking

Did you know there are self-driving vehicles? They rely on AI to ensure they don’t crash
Have you ever played a player vs player game and played against a computer? Whether it’s tic-tac-toe, chess, or Nintendo Super Smash Bros AI is probably used to help the computer decide it’s next move

Creating

Have you ever used a tool that creates new images based on a sentence you type?
Have you ever used a tool that writes a summary of information for you, or writes a story or email?

ChatGPT and these types of tools are also AI

Related articles

8 of 38

8

HOW DOES AI WORK?

AI looks at data from the past

Analyzes the data to find patterns

Uses the patterns to make predictions

Instructor Dialogue

“How does AI work?”

First you have to provide the AI model with lots and lots of data. What type of data depends on what you are trying to teach the AI.

AI analyzes the data to find patterns and then

AI uses the patterns to make predictions.

Instructor Cues

Here are specific examples you can share

First you have to provide the AI model with lots and lots of data. What type of data depends on what you are trying to teach the AI.

If you are trying to teach an AI to understand a person talking, you would give it lots of recordings of people talking and transcripts of what was said in each recording.

If you are trying to teach an AI to make up a story, you would give it lots of different stories

If you are trying to teach an AI to play tic-tac-toe, you would give it the moves from lots of example tic-tac-toe games and the results of who won or lost each game

AI analyzes the data to find patterns

The AI learning to understand a person talking might notice that whenever it hears a “ssss” sound that there is a word containing a c or s

The AI learning to make up a story, might notice most stories have a happy ending

The AI learning to play tic-tac-toe might notice that the player with the middle square in tic-tac-toe wins more games

AI uses the patterns to make predictions

The AI learning to understand people hears a new word with a “ssss” sound and an “oh” sound and guesses the word is so (of course it might be wrong maybe the word is sew!)

The AI learning to make up a story ends a story with the main character winning a race

The AI learning to play tic-tac-toe places it’s piece in the middle square if that spot is available

9 of 38

9

WHAT IS AI?

Term	Definition
MODEL	�An AI Model is a program that has been trained using data to recognize patterns and make decisions or predictions. Each AI tool uses an AI model.�

10 of 38

10

ARE YOU AS SMART AS AI?

11 of 38

11

HOW MANY SECONDS WILL IT TAKE TO WALK 5 METERS?

Distance walked	1 meter	2 meters	3 meters	4 meters	5 meters
Number of seconds	2	4	6	8	10

?

12 of 38

12

DID YOU SPOT THE PATTERN?

Seconds

Meters

One of the ways to look for a pattern is to plot the data on a graph

In AI, when the dots form a straight line, the pattern is called linear regression.

Nbr Seconds = Nbr Meters * 2

13 of 38

13

REGRESSION HELPS PREDICT HOW MUCH ARCTIC SEA ICE

MELTS FROM ONE MONTH TO THE NEXT

Arctic sea ice expands and thickens in winter and then shrinks and thins in summer. This affects shipping routes as well as the lands and lives of Arctic peoples. What factors might affect how much ice melts?

Factors considered in AI models include

Ice thickness
Temperature
Surface area of the ice

Instructor Dialogue

“Arctic sea ice expands and thickens in winter and then shrinks and thins in summer. This affects shipping routes as well as the lands and lives of Arctic peoples. What factors might affect how much ice melts?

Factors considered in AI models include

Ice thickness
Temperature
Surface area of the ice”

Instructor Cue

Give the students a chance to share their thoughts on what factors might affect how much sea ice would melt. They may suggest other factors not listed here. There is a lot of research going on trying to predict arctic sea ice melt using many different forms of AI and many different measurements to help with predictions.

The intention here is to learn that when you are trying to make predictions if there are multiple factors that affect the outcome, then when you train an AI you want to provide data on as many of those factors as possible to get the best predictions

Related articles

Artificial intelligence to help predict Arctic sea ice loss

�

14 of 38

14

WHERE SHOULD THE KANGAROO GO?

GROUP B

GROUP A

15 of 38

15

DID YOU SPOT THE PATTERN?

Group A contains plants

Group B contains animals

When AI uses patterns to divide objects into groups that is called classification

16 of 38

16

WHERE DOES A COFFEE MUG GO?

ANIMALS

PLANTS

17 of 38

17

AI WILL ALWAYS ASSIGN IT TO A GROUP

ANIMALS

PLANTS

18 of 38

18

HUMANS MIGHT DECIDE WE NEED A NEW GROUP!

ANIMALS

PLANTS

OTHER

19 of 38

19

CLASSIFICATION COULD BE USED TO SORT TRASH

20 of 38

20

IF I ROLL A DICE 5 TIMES, WHAT NUMBER WILL I ROLL NEXT?

Rolling a dice returns a random number, if you get a pattern it’s just luck

Even AI can’t predict what the next number you roll will be, but it will try

21 of 38

21

HOW DO I KNOW HOW MUCH TO TRUST THE AI?

AI will always make a prediction even if it has trouble finding a pattern

AI tools may return a ‘confidence level’ from 0 (not confident at all) to 100 (very confident) that indicates how confident the AI is in it’s prediction.

I am 23% sure it will roll a 6.

22 of 38

22

HOW DO YOU TEACH AI TO FIND CATS IN PHOTOS?

You give the AI lots of photos.

You tell the AI which photos have cats.

You tell the AI where the cats are in the photos.

The AI identifies patterns.

For example: Cats are usually furry, with 2 eyes, pointy ears and a tail

When you give the AI a new photo it uses that pattern to find cats.

This is called object detection.

CAT

NO CAT FOUND

Pointy ears

Tail

2 Eyes

Furry

CAT

NO CAT FOUND

Instructor Dialogue

“Let’s look at how you would train an AI to find cats in photos

First you have to provide the AI model with lots and lots of data - in this case that would be pictures of cats and pictures that do not contain cats

Since we are teaching our AI to find cats we need to give it lots of pictures of cats
It’s important to teach the AI which pictures do NOT contain a cat as well, so we also give it pictures of other animals
Every time we give the AI a photo we identify any cats

AI analyzes the data to find patterns in the photos of cats

The AI compares the features of the animals we identified to see what cats have in common
The AI might notice a pattern that in most of the cats in the photos have tails, pointy ears, 2 eyes and are furry.

AI uses the patterns to identify photos that contain cats

Once the AI identifies a pattern (tail + pointy ears + 2 eyes + fur = cat) you can give it new photos and if it sees that pattern in the new photo it says here’s the cat!

23 of 38

23

IS THERE A CAT IN THIS PHOTO?

To fix it, you would have to retrain your AI using dog photos so the AI can find a new pattern that can distinguish between cats and dogs.

The quantity, variety, and quality of data affects how often an AI is right or wrong.

CAT

The dog matches the cat pattern so the AI thinks it found a cat!

Pointy ears

Furry

2 Eyes

Tail

Instructor Dialogue

“Is there a cat in this photo?

You are smarter than the AI! Based on the pattern the AI learned, it says since I see something with 2 eyes, pointy ears, a tail and fur this is a cat!

To fix the AI and make it ‘smarter’ , we would have to retrain the AI with lots of pictures of dogs so it can learn a new pattern that allows it to tell dogs and cats apart. It might learn that cats have whiskers, or that cats are usually in different poses than dogs.

The more data you provide, and the more variety you provide, and the quality of the data (e.g. higher resolution photos where you can clearly see the animal) the more likely your AI will be able to identify a pattern to correctly make predictions.”

Instructor Cue

Give the students an opportunity to answer the question before you use the slide animation to reveal that the AI will say this is a cat

The goal of this slide is to demonstrate that AI will make mistakes and that often humans, even kids, are usually smarter than AI!

24 of 38

24

WHICH OF THESE PHOTOS CONTAINS A CAT?

Would your answer change if you were training the AI to

Monitor endangered species visiting a watering hole in Africa?
Use city traffic cameras to help find lost pets?

Each AI tool is trained for a specific purpose.
Before using any AI tool, research what it is designed to do and how it was trained. Use the right tool for the right job!

Instructor Dialogue

“Which of these pictures contains a cat?

It depends on your definition of cat! When you are training an AI how you label the data (in this case whether the photo is a cat or is not a cat) depends on the purpose of the AI

If the AI is being trained to help with monitoring endangered species visiting a watering hole in Africa, you would label the photo of the cheetah as a cat, but the toy and the pumpkin photos would be labelled as photos that are NOT cats
If the AI is being trained to use city traffic cameras to help find lost pets, none of these would be cats they would want to rescue so you would label all the images as photos that are NOT cats
If the AI is being trained to do a search through a photo library for images of cats, you would probably want all these images returned, so you would label all these photos as photos that contain cats

If you use an AI, and you don’t know its purpose or how it was trained. You may get undesirable results because your definition or expectation is different from the person who trained the AI.

If you try to use an AI trained to help the humane society find lost pets to idenitfy lions and tigers from wildlife cameras in Africa it won’t work very well because it wasn’t trained to identify lions and tigers as cats.

�Does that mean the AI is making mistakes and is wrong? No, it just means you have selected an AI tool that is not well suited to your task.

Research the AI tool before you use it so you can use the right tool for the right job!”

Instructor Cue

There is a good chance your students will disagree on which images contain cats.

Ask individual students why they say a particular photo is or is NOT a cat. Everyone who gives an explanation is right. If everyone in the class was training their own AI, each of their AIs would make the predictions based on THEIR definition of a cat.

Then reveal the text below the photos and see if they agree on which photos contain cats for the specific scenarios.

Related articles

Using the power of AI to identify and track species

25 of 38

25

OBJECT DETECTION IS USED TO TRACK FISH SPECIES

The challenge: You can’t know if you have a sustainable fishery if you don’t know how many fish are coming back.
The solution: An underwater camera in Indigenous-run fish weir and an AI model identifies 12 different species of fish with 80-90 % accuracy to help track salmon populations.

Instructor Note

Pick ONE Object Detection example to share with the students, share the example that is more likely to resonate with the students either the Beluga monitoring with aerial images or the underwater camera for tracking fish in a river.

Instructor Dialogue

“Object detection is used for fisheries management. In rivers across British Columbia’s Central Coast we don’t know how many salmon are actually returning. At least not until fishing seasons are over. Meanwhile fisheries managers still have to set harvest targets for commercial and recreational fisheries. If they get it wrong they have to make emergency closures. With accelerating climate change, it’s becoming more and more unpredictable so we can’t assume the patterns will be the same from one year to the next. Salmon Vision uses underwater cameras to identify salmon swimming through Indigenous-run fish counting weirs. Now we can monitor the salmon population in real-time helping track the species caught by First Nations, commercial, and recreational fishers.

https://wildsalmoncenter.org/2023/10/18/artificial-intelligence-meet-indigenous-fishing-technology/

Related articles

Artificial Intelligence, Meet Indigenous Fishing Technology

26 of 38

26

OBJECT DETECTION IS USED TO TRACK BELUGA WHALES

AI Image analysis can identify beluga whales with 85-88% accuracy using aerial, drone and satellite imagery.
The model can analyze thousands of images in hours, a task that previously took months to complete.

27 of 38

27

PICK THE BEST AI TOOL FOR EACH JOB

As you can see there are a lot of different types of AI

It’s important to pick the best AI for an individual task to get the best results

28 of 38

28

WHAT IS GENERATIVE AI?

Term	Definition
GENERATIVE AI	�AI that creates new content such as text, image, music, or video. They create content based on patterns in the data used to teach the AI�

29 of 38

29

WHICH OF THE EXAMPLES BELOW IS GENERATIVE AI?

Your video app recommends movies

You ask Meta (Facebook) to imagine a photo of a cat riding a flying turtle

You ask ChatGPT to make birthday card greetings for a soccer player

Recommended movies

30 of 38

30

HOW DOES GENERATIVE AI CREATE ORIGINAL TEXT?

Term	Definition
LARGE LANGUAGE MODELS (LLM)	�LLMs analyze vast amounts of data to find patterns so they can generate and respond to human language in a way that mimics natural conversation�

31 of 38

DOES GENERATIVE AI MAKE MISTAKES?

If your teacher asked Generative AI to write your report card it might have analyzed report cards in other countries including the United States.

Students in the United States usually have American history as a subject on their report card.

So the generative AI might add American history to your report card, even if you never took that subject.

You should ALWAYS fact check content created by generative AI

AMERICAN HISTORY

YES!

Instructor Dialogue

When you ask generative AI “What is 2X2?” , it says 2X2 = 4 because in the past it has seen a lot of examples of equations and most of the time when it saw 2x2 it saw the answer was 4. Calculators do math to calculate results of equations. Generative AI just says 4 is what I usually see as an answer so I’m going to say 4 is the answer because it matches the usual pattern.

Imagine if your teacher asked generative AI to write your report card. The generative AI was trained by analyzing other data including other report card. Some of those report cards are from other countries including the United States. Students in the United States usually have American history as a subject on their report card so the generative AI might add American History to your report card, even if you never took that subject!

You should ALWAYS fact check content created by generative AI.

Instructor Cue

You can also ask the students if just like the image model that was trained to identify cats, is there a risk an LLM will make predictions we don’t expect or predictions we consider wrong because we don’t know exactly how the model was trained.

The answer is YES! If the only data used to train a large language model is Shakespeare plays and sonnets, the LLM is going to suggest very different words than an LLM trained using nothing but text messages!

32 of 38

DOES GENERATIVE AI MAKE MISTAKES?

Term	Definition
HALLUCINATION	�When AI generates incorrect, fabricated or nonsensical information that appears plausible but is not grounded in reality�

33 of 38

33

AI HAS AN IMPACT ON THE ENVIRONMENT

Training a large language model can consume as much energy as 100 homes consume in a year
By 2026, it’s estimated 6% of the total energy consumption in the U.S. will be for AI
Fresh water may be pumped in to cool server rooms and pumped back to the stream or lake at a higher temperature

Instructor Dialogue

“AI has an impact on the environment.

With all the excitement around AI, sometimes we forget about potential negative consequences.

Training a large language model can consume as much energy as 100 homes consume in a year!
By 2026, it’s estimated 6% of the total energy consumption in the U.S. will be for AI

To minimize the impact of AI on the environment

Use models trained for specific tasks wherever possible, they consume less energy
Just like water and electricity, don’t be wasteful using AI

If you want to learn more, you can find links to articles about AI and the environment in your Student Workbook Appendix A - AI and the Environment”

Instructor Cue

If you want to learn more, here are some articles with sample quotes related to AI and the environment you can share with your students

AI and Sustainability: Opportunities, Challenges, and Impact by Dr. Volha Litvinets Senior Digital Risk Consultant, Responsible AI, Ernst & Young Advisory. This article explores the environmental impact of artificial intelligence (AI) and discusses opportunities, current legislation, and ways organizations can make AI usage more sustainable.

The Uneven Distribution of AI’s Environmental Impacts by Shaolei Ren and Adam Wierman - Harvard Business Review

“the training process for a single AI model, such as a large language model, can consume thousands of megawatt hours of electricity and emit hundreds of tons of carbon. This is roughly equivalent to the annual carbon emissions of hundreds of households in America”

“In the United States, the rapidly growing AI demand is poised to drive data center energy consumption to about 6% of the nation’s total electricity usage in 2026”

Silicon Valley and the Environmental Costs of AI - Ana Valdavia - Political Economy Research Centre

“Designing an AI-based algorithm does not require particularly expensive infrastructures. But if these algorithms are trained with large amounts of data or the algorithmic architecture is sophisticated, then it is necessary to use servers with greater computational capacity to store and process the information. To build these chip-based infrastructures, raw materials are necessary: iron, aluminium, minerals such as lithium, silicon or cobalt and water, among others.”

“AI does not rely on raw materials only during the construction of its physical infrastructures; it does so throughout its cycle. For instance, data centres and servers need large amounts of water to cool down. According to a study published in Nature in 2021, Google and Microsoft declared using respectively 15.8 billion and 3.6 billion litres of water.”

“The more data to be analysed, the higher the energy consumption. More sophisticated algorithms, which need long computational time, consume even more. For example, it is estimated that training an algorithm to automatically produce text uses 190,000 kWh; that is, 120 times more than the average annual consumption of a household in Europe in 2020”

Generative AI’s environmental costs are soaring and mostly secret - Nature

“one assessment suggests that ChatGPT, the chatbot created by OpenAI in San Francisco, California, is already consuming the energy of 33,000 homes.”

“It’s estimated that a search driven by generative AI uses four to five times the energy of a conventional web search”

DataBack Vol 2: Truth Before “Indigenous AI” Understanding and Resisting Generative AI - Animikii Inc.

“the ongoing energy consumption and carbon emissions from AI inferences — tasks where the model is applied to real-world data — are significantly higher for multi-purpose AI models compared to task-specific models”

“multi-purpose models consume orders of magnitude more energy than task-specific models.”

“The more we use Generative AI, the worse the impact is on the environment”

34 of 38

34

DOES THAT MEAN WE SHOULDN’T USE AI?

In fact, there are ways to use AI that help the environment

After a lightning storm you can send drones to fly over forests and use AI to analyze images to locate forest fires and put them out quickly before they spread

NO!

35 of 38

35

BE THOUGHTFUL WHEN YOU USE AI

When choosing an AI, look up the sustainability efforts of the provider, are they making efforts to minimize their environmental impact?
Wherever possible use an AI designed specifically to do the task you want to accomplish. AIs designed to do specific tasks use less energy e.g. face ID on your phone vs ChatGPT

36 of 38

36

WHAT AI WOULD YOU LIKE TO CREATE?

?

37 of 38

37

NOW YOU KNOW!

Key Takeaways

AI is all around us
AI makes predictions based on patterns and can make mistakes
We need to be thoughtful about using AI

38 of 38

38

WANT TO KEEP LEARNING ABOUT AI?

Get hands-on and learn more about Generative AI and in our

“What is the role of AI in Songwriting? workshop!”