To complete today’s task I need to create an image of myself based on my job title and also create an image of myself based on my role in society.
Although Adobe Firefly, DALL-E or even Midjourney, Stable Diffusion, Starryai, Dream by WOMBO, NightCafe or ArtBreeder was recommended to create the images, I used Image Creator in my Microsoft Bing account to create the images, which is powered by powered by DALL·E 3.
Asking Image Creator to create an image based on my job title
Using the prompt “A weary middle aged man, cropped hair, no beard in a white shirt black tie and rolled-up sleeves working with intent at a computer in a sea of technology and kaiju designing digital learning for everyone everywhere around the world, in the style of black and white manga”, Image Creator created an awesome and entirely job-accurate image. Impressive.
Asking Image Creator to create an image based on my role in society
Using the prompt “A weary middle aged man with cropped hair, no beard in a white shirt black tie and rolled-up sleeves and bass guitar on back and skateboard in hand defending his psychologist wife, gamer teen son and punk tween daughter from a horde of evil technology and crazy kaiju and raining hellfire, in the style of black and white manga”, Image Creator created another awesome and entirely societal-role image of me and my family.
Reflecting on the Day 3 task
My experience of prompt-based image creation was not dissimilar to other learners on Day 3.
How’d I get on with the task? I’ve had some experience with AI image creation before e.g., Midjourney and some other tools made available through Discord, so it wasn’t an entirely new to me, but still fun as always. And, because I didn’t want to create a photo-real image of myself for the task I leaned into more stylistic representations.
What’d I think of the generated images? I was happy with them and had fun with applying a “in the style of…” type filter to modify the image. I’d be interested in learning how to prevent random words from appearing as part of the image, if that’s possible – I’ve heard running a local installation of Stable Diffusion permits greater control over the image format as well as what’s included and excluded in the generated image.
Were the generated images surprising? Not really. Image Creator pretty much generated what I asked for, more or less.
Is AI stifling creativity – or allowing room for more to flourish? For me, it’s allowing room for more people to flourish and express themselves. Like any tool it’s an enabler and you get out what you put in – prompt engineering is a skill like any other.
What about any bias or stereotypes in the images? Nothing that wouldn’t be expected from a prompt that specified a “black and white manga” style. With that being said, I wouldn’t be startled to learn that bias was evident in images created to be “more realistic” or “photo like”, as noted in Humans are biased. Generative AI is even worse.
Was there a need to modify the prompt to make it more representative of the job/position I was looking for? Yes, to ensure some additional elements were included in the generate image, which makes sense.
How might these tools be useful in my professional practice? Like ChatGPT and Claude or any other tool, AI powered image creation tools can be considered a starting point or to help kick-start the creative process.
On Day 3 there’s an interesting note on how “things can get interesting when we start combining AI tools together. Earlier this month as part of our AI Conversations series, Dave White (XYZ, UAL) discussed answers to the question “Is AI stealing our Creative Labour?” After the session, Chris generated an AI-written blog post on the transcript using ChatGPT, Claude and Adobe Firefly. Read the post on his blog“. Great stuff.
Also on Day 3 are some recommended optional readings:
- AI and the Rise of Mediocrity (TIME Magazine)
- This is how AI image generators see the world (Washington Post – requires free account to read)
- Humans are biased. Generative AI is even worse. (Bloomberg UK)
And for no particular reason other than fun, I asked Image Creator to create an image of “itself” painting a picture of me.