‘Mama Day’: AI, Art & Afrofuturism
I’m kicking off Black History Month, also referred to as Black Futures Month, with Black women sci-writers. I’m also preparing to give a talk next week on “AI, Art & Afrofuturism,” so the theme is timely.
Rooted in themes of science fiction and fantasy genres, [Afrofuturism] sits under the umbrella of the speculative fiction subgenre, as well as, recontextualizing the black identity, placing black people in a positive light as leaders, innovators, & influencers of societal change. — Mark Dery, “Black to the Future” 1993
At the intersection of AI, art and Afrofuturism is visual storytelling, or what I refer to as “visual storytelling 3.0.” This aligns with recent developments in generative AI such as the use of images and text as prompts for AI-generators. Visual storytelling 3.0 is driven by variation and remixing, among other modes and options. For example, the variation buttons in Midjourney allows me to change the composition, number of elements, colors, and the type of details within an image.
The current version of Midjourney or MJ v6 can understand prompts more accurately and it has improved image prompting and remixing. Remix mode lets you change prompts, model versions, or parameters between image variations. When Remix is enabled (see above), it allows you to edit your prompt during each variation. For example, I wanted to create a portrait of writer Gloria Naylor who wrote Mama Day, one of my favorite novels. Mama Day employs the literary technique called “magical realism,” in which elements of dreams, fairy-tales, and mythology are combined with recognizable everyday reality.
Author Ytasha Womack lists multiple sub-genres around Afrofuturism including “science fiction, historical fiction, speculative fiction, fantasy, Afrocentricity, and magical realism (with non-western beliefs).”
To begin the process, I used a simple text prompt: “golden Gloria Naylor.” I wasn’t sure what the tool would produce and I was pleasantly surprised because the images looked a little like Naylor, including this one:
Next, I used the Pan option that allowed me to expand the canvas of the Naylor image in a down direction without changing the content of the original image. I subtly upscaled the image and saved it to my computer. While the initial image looked somewhat like Naylor (in the ballpark) it wasn’t close enough, so I edited it using Adobe Photoshop. I changed the face, aligned and reshaped the eyes and eyebrows. I also added one of Naylor’s cheek dimples.
I went back to MJ v6 and used Remix mode to add a specific quote from Mama Day (see above). I did not re-interpret it, as usual. I wanted to incorporate some of the magical realism into the portrait and the quote I chose was very visual. I ran the generator a few times, upscaled a couple and composited everything in Photoshop to create this revised image:
In the intermediate, remixed version of the AI-generated Naylor portrait, she is grabbing lightning and, using Photoshop, I composited in the moon and stars on the right side of the image. I also subtly edited the hands. As a final step, I remixed the original image again to add the “medicine bowl.”
Visual storytelling helps people make sense of complex data, elicits emotion and increases information retention. Using the different modes and options in Midjourney v6 allows me to stitch together or composite multiple AI-generated images to pare down the text to the bare essentials and let the visual speak for itself.
I have a dream for the Web [in which computers] become capable of analyzing all the data on the Web — the content, links, and transactions between people and computers. A “Semantic Web”, which makes this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The “intelligent agents” people have touted for ages will finally materialize. — Tim Berners Lee on the “Semantic Web” in 1999
As a final step, I took a part of the selected quote and used it in a MJ v6 remix, then composited it with the previous version in Photoshop. The result captures the vibe or feel of the Naylor text, while also portraying the late author as if she stepped into her own story. This is visual storytelling 3.0 because it taps into the Semantic Web — a web of data that is processed by a machine such as generative AI — in order to visually manifest some of the magical realism in Mama Day.