Create drawings by hand for a given prompt. Compare your drawings with others. Discuss how each one draws differently for the same prompt.
Click here to open the instruction document and sample prompts for this exercise.
Exercise (2) - Image generation tool (Gemini)
Part 1: Use an AI image generation tool like Fotor to create images from different prompts. Observe how adding more details to the prompt changes the generated image.
Part 2: Generate images using Gemini and give related prompts to observe how AI retains context. Discuss how previous prompts influence the generated images.
Click here to open the sample prompts for this exercise.
Exercise (3) - Wild prompting
Use your imagination to come up with a wild prompt such as 'Dinosaur wearing a spacesuit on Mars'. Compare AI-generated images with Google search results using the same prompt for both. Discuss how AI creates new images even when search fails to look up such images.
Click here to open the sample prompts for this exercise.
Use Gemini or similar image generation tools for Exercise 2.
Emphasise key ideas such as context, keywords and prediction.
Discuss the similarities and differences between how we imagine and draw images vs how AI does.
Explain that AI tools guess patterns based on the prompts / text clues / keywords and then generates the closest matching patterns.
Highlight the difference between images available on the internet and AI-generated images.
Help students with sample prompts linked in each exercise if they are unable to come up with their own prompts
Points to Ponder:
Everyone was given the same prompt, but no two drawings came out the same. Why? Does AI give a different picture each time for the same reason you did?
Google finds a picture that already exists somewhere. AI makes one that has never existed. How do you think AI is creating images that it has never seen before?