1. Examine images.
a) The following images are generated with the AI software Stable Diffusion. They have been generated using the text “An old woman crying, portrait, photo, winner, stunning composition, beautiful light”. What kind of culture do the images represent? (Images: Tomi Knuutila)
b) Perform a Google image search with the phrase “old woman crying”. How do the images differ from the ones generated with the AI and why? You can also use the following image collage from a performed Google search.
c) Compare two images. One of the images is a photograph from the internet and the other is generated with the AI software Stable Diffusion and enlarged with the AI algorithm ESRGAN. Is it possible to say based on the image whether it has been generated with an AI or whether it is taken by a person? Based on what?
Image 1.
Source: https://www.flickr.com/photos/worldbank/3426922863
Image 2.
2. Create images. Come up with a text-based description that you would like to enter to an image-generating AI. Discuss what kinds of images might be generated based on the text. Try to enter the suggested texts with your teacher to the AI at https://creator.nightcafe.studio/ or https://dream.ai/create. The texts have to be translated into English. Did the images meet your expectations? What if you would enter the texts again?
3. Come up with your own AI: Divide the students in small groups and ask them to come up with their own AI. What kinds of tasks their AI could perform? How would it work? This assignment might motivate the students to think about what kinds of problems an AI could solve.
Split into small groups and come up with your own AI. What kinds of tasks your AI could perform? How would it work? This assignment might motivate you to think about what kinds of problems an AI could solve.
One of the assignments is created by Chat GPT. Can you recognise which of the assignments is by an AI?
Text: Tomi Knuutila
