Meeting 6
Meeting 6
For the virtual background, use one of the image generators linked below to make a virtual background of the location you wish this class was held (or whatever else you want). Stable Diffusion lets you use 16:9 aspect ratio, which is good for a virtual background. (Or if you don’t want to generate one, you can use one of the virtual backgrounds from a previous week.)
First Activity: AI Image Cliches
I have posted a Google drive (link removed) with almost all the images submitted for assignment 6 last quarter. (View in “grid layout” to see many images quickly while scrolling, or open individual images full screen and use arrow keys to move back and forth.)
Go to the Google Sheet (link removed) for this assignment, and find the tab for your group.
Look through the images that were submitted. If you see a recurring cliche, add it to your sheet, and list as many images as you can find that have it. I’ve given a few examples on the front tab.
- Are there people/robots in identical poses despite the image otherwise being different? Is the same color used in the same way? Is there a bit of background detail that’s the same? Are there recurring facial expressions? Is there an art style that keeps coming up? Is there a particular pattern of lines or pattern of colors that you see the same in very different images?
As you’re looking through, if there are images you really like, add them to the cell for your favorites.
There’s too many images to look through all of them full screen - but if you are in Group 1, make sure you at least look through images 0100 through 0199, if you’re in Group 2, make sure you at least look through images 0200 through 0299, and if you’re in Group 3, make sure you at least look through images 0300 through 0399.
One of the tabs is an “Image Key”, which says which fiction and which AI system each image was from (according to the homework submission) in case you’re curious about any of them, and so you can see if some of your “cliches” are actually elements of a specific fiction, or quirks of one of the AI systems.
After spending about 20-30 minutes doing this, look through the other tabs and see if other groups have found similar cliches or different ones.
Second Activity: AI Art Turing test
Get into groups of 3-5. One person should open this AI Art Turing Test and share their screen - you should all look at the images together and see if you can agree whether it’s human or AI art, and explain why you think that, or if it’s just a vague sense. See if you’re picking out the same features, see if certain types of art are easier to identify, see if you’re just getting it wrong.
Afterwards your group should also try at least five faces on https://www.whichfaceisreal.com/index.php.
Third activity: Discuss this module’s ideas
Open up the Canvas page for module 6 (link removed) to remind yourselves of the videos for this module. Discuss these videos with each other.
Were there any ideas from any of them that you found particularly surprising, or interesting? Were there things you didn’t understand?
Look at the list of “Further watching/reading” links below the actual assignments. Are there some that look interesting to dig into further? Are there any you already looked at that some of the other people in your group might want to check out?
Free AI image generators
One good free AI image generator is Stable Diffusion. You might need to put an e-mail for a login - I recommend using a throwaway e-mail account, or creating a new e-mail account if you’ve already used all your free generations for the day. I recommend using “Number of Images: 1” to save your free tokens for the day, and pay attention to “Aspect Ratio” if it’s relevant.
If you don’t want to come up with your own prompt, try this Image to Prompt site.
You can also use ChatGPT or Gemini for some number of free image generations per day - you don’t need to give as much detail in your description of the image (these LLMs have a lot of background knowledge) but you also don’t get as much control over the image. Comparing the results you get from Stable Diffusion, ChatGPT, and Gemini will likely also be interesting. (And I have a subscription to Midjourney, so feel free to call me over to your room if you want to see another version.)