Now that you’ve learned about what fuels AI, it’s your turn to become an AI engineer! In this activity, you will make decisions to build a sustainable AI model.
Use the Scratch game below to build a high-performing AI model with minimal environmental impact. To do this, you’ll need to choose the right hardware, dataset, cooling method, and energy source.
How to Play:
- Start the Game: Click the green flag to begin.
- Select Your Resources: Click on the cards to select a choice for the Hardware, Dataset Size, Cooling Method, and Energy Source variables.
- Calculate Your Impact: Once you’ve made all four choices, select ‘Calculate Impact’.
- Analyze Your Results: Review the bar graph in the bottom left of the game to see how your choices have impacted the AI’s performance and environmental footprint.
- Try Again! Adjust one choice at a time until you find the best balance across AI performance and environmental footprint. Can you explain why some combinations are more sustainable than others?
LET’S REFLECT…
After you’ve run the simulation and reviewed your final graph, take a moment to consider your results.
- Which choice had the biggest impact on the outcome?
- How did your choices balance AI performance with energy consumption?
- In your opinion, is it possible to select a ‘perfect’ combination? Why or why not?
- Are there any other choices or variables you wish were captured in this simulation?