Home AI for Earth's Future
🌍 Module 04 · Advanced

AI for Earth's Future

What if you could be a high-tech detective for the entire planet? 🕵️‍♀️ Not just for weather, but for protecting animals, oceans, and forests! AI is a powerful tool that helps us understand and protect our world. You'll even train your own AI with your webcam to sort recycling—no code needed! Let's get started. 🚀

🛡️ Safety Check: Handling Eco-Anxiety

Learning about environmental challenges can sometimes feel big and scary. That's a normal feeling called 'eco-anxiety'. But remember: every scientist, engineer, and builder you see here started just like you—curious and wanting to help. Focus on the one small, cool thing you can learn or build today. That's how we build a better tomorrow, one step at a time.

🐅

AI: A Guardian for Wildlife

See how AI helps conservationists analyze millions of hidden camera photos to protect animals. (Video: ZSL)

From tracking whale songs across the ocean to spotting poachers from space, AI is giving conservationists superpowers! It can analyze gazillionsWhoa, That's a Lot of Data! 🐘
Scientists use a word called a 'petabyte'. It's so much information that if it were a movie, it would take you over 200 years to watch! That's why we need super-fast computers to help.
of data points—like photos, sounds, and satellite images—to help protect endangered species and their habitats.

🧠 Data Detective Challenge: Spot the Trend

An AI model's first job is to find patterns. Look at this simple climate data. First, check out the chart—you can *see* the trend! Then look at the numbers. Can you spot the pattern a computer would see?

2020
2021
2022
2023
    Year: 2020, Avg Temp Anomaly: +0.98°C
    Year: 2021, Avg Temp Anomaly: +1.05°C
    Year: 2022, Avg Temp Anomaly: +1.12°C
    Year: 2023, Avg Temp Anomaly: +1.19°C
          

Question: Based on the trend, what is the most likely temperature anomaly for 2024?

A) +1.15°C (Cooler)
B) +1.26°C (Warmer)
C) +1.50°C (Much Warmer)

🚀 Challenge Question:

If this trend of +0.07°C per year continues, in approximately what year would the anomaly first pass +1.50°C?

A) 2026
B) 2028
C) 2030

Thinking Like a Computer: Step-by-Step

Before we use a fancy code shortcut, let's think about how you would find the hottest year without one. You'd probably:
1. Grab a notepad and write down the first year's temperature (+0.85°C).
2. Look at the next year (+0.98°C). It's bigger, so you erase the old number and write down the new one.
3. Keep doing this for the whole list.
This step-by-step process of comparing and updating is exactly how a computer "thinks"!

🤖 Teaching the Computer to Think

You just spotted the trend like a real data detective! Awesome. But what if you had 1,000 years of data? That would take forever. This is where code comes in. We give the computer an instruction like this:

Find the BIGGEST [temp_anomaly] in the [climate_data] list.

Now let's see how that looks in Python code.

Code Breakdown

1
The Data: climate_data = [...] is like a digital filing cabinet. The [] means it's a list, and each {} is a file folder with labeled data inside.
2
The Superpower (Shortcut!): max(...) is a built-in Python function that does all the 'notepad' work we just described automatically. It loops through the list and keeps track of the biggest value it has seen so far.
3
The Target: key=lambda item: item['temp_anomaly'] is how we aim the superpower. We're telling max(): "Don't just look at the year or the CO2. I only want you to compare the numbers labeled 'temp_anomaly' to find the biggest one!"

Python Sandbox: Find the Hottest Year

# Complete the code by filling in the blank!

Awaiting analysis...

🚀 Advanced Challenge: Above Average

Climate scientists rarely look at a single data point. They look for trends. Your mission:
1. Calculate the average temperature anomaly for all the years.
2. Write a loop that prints out only the years that were hotter than the average.

Advanced analysis awaiting your command...

💡 What If The Data Is Goofy?

Real-world data is messy! What if a weather balloon pops and sends back a crazy number? An AI needs to be smart enough to handle it. What should it do?

Question: A sensor sends back a temperature of 2000°C! What should our AI do?

A) Use the number, our planet is doomed!
B) Guess that the sensor is broken and ignore that one number.
C) Stop working forever.

Awesome work! You've taught the computer to understand patterns in numbers. But what about what computers can see? Let's switch from analyzing data tables to analyzing images. The AI thinking is the same: find patterns!

Hands-On Project: Build a Smart Recycling Sorter

Enough code for now—let's build a real AI! We're going to use a super cool tool called Teachable Machine from Google. It lets you train an AI model right in your browser, no code required. Your mission: teach it to tell the difference between paper, plastic, and metal.

See how Teachable Machine works in just 2 minutes! (Video: Google AI)

1
Gather Your Materials: Find a few recyclable items around your house. Grab a piece of paper (a receipt or junk mail), a plastic bottle, and an aluminum can.
2
Open the Tool: With a parent's permission, click this link to open Teachable Machine. Click "Get Started" and choose "Image Project."
3
Train Your AI: You'll see different "Classes." Rename Class 1 to "Paper." Hold up your piece of paper to the webcam and click and hold "Hold to Record" to take lots of pictures. Do the same for "Plastic" and "Metal" in the other classes. The more pictures, the smarter your AI will be!
4
Test It!: Click the "Train Model" button (this might take a minute). Once it's done, hold up your items to the webcam again. Does the AI correctly guess what each item is? Try crumpling the paper—does it still know?
5
Take It Further: Export Your Model. Notice the "Export Model" button after you train your AI. Click it! You'll see code snippets for Javascript (p5.js) and Python (Tensorflow). This is how a real developer would take a model from a training tool and use it in a real app. We'll explore this in our Web and Game Development tracks!

✍️ Your Turn: Be an AI Designer

You're a conservationist trying to protect the endangered Snow Leopard. You have thousands of camera trap photos but no time to look through them all. Your mission is to write a "job description" for an AI assistant. What tasks would you want it to do? What skills would it need?

Example Task: "Scan all new photos and alert me *only* if a snow leopard is identified."
Example Skill: "Must be able to tell the difference between a snow leopard and a fluffy sheep."

Think of two more tasks and two more skills for your AI!

🛡️ Safety Check: Thinking About AI Bias & Privacy

1. AI Bias: What happens if your AI only ever sees clean, perfect plastic bottles? Will it recognize a crumpled, dirty one? This is called "bias"—when an AI is only trained on one type of data, it gets confused by new things. Real-world AIs need to be trained on LOTS of different examples!

2. Your Face is Data: When you use your webcam, you're creating data. Only use tools like Teachable Machine that don't save your personal data, and always ask a parent before turning on your camera for any website or app.

👨‍👩‍👧 Parent Corner: Talking About AI

The Teachable Machine project is a fantastic opportunity to talk about how AI works in the real world. You can discuss AI ethics with questions like: "If an AI is sorting recycling, who is responsible if it makes a mistake?" or "What kind of jobs could an AI like this help with?" This shifts the conversation from "magic" to a powerful tool people design and manage.

📚 Go Deeper

  • NASA's Global Temperature Data - See the real, live data that scientists use every day.
  • Kaggle Climate Datasets - Ready for a real-world data mission? Challenge: Download the 'GlobalLandTemperaturesByCity' file. Can you use a tool like Google Sheets or Python with the Pandas library to find the average temperature for your city in the earliest and most recent years in the dataset? (Requires parental supervision to create a Kaggle account).