“WRDS & Images:
Intersections of writing, rhetoric and digital studies”
by Justin R. Cary, Senior Lecturer in the Writing, Rhetoric & Digital Studies Department, The University of North Carolina at Charlotte
A picture is worth a thousand words.
This cliche echoes through history as an affront to the power and impact of writing; the worth of a picture certainly cannot be measured in a simple word count. And the worth of writing cannot be measured in a single image. Of course, this statement should not be taken literally but simply as an implication of the connection between words and images, the intrinsic link between alpha numeric text and visual idea representation in an image; signifier and signified.
The tree came first; trees existed long before humans developed hierographic and language systems to encode the the tree into a picture sketched roughly on a cave wall. Later, we discovered a way to encode the tree into letters but those four letters in that arbitrary, socially agreed upon order have nothing ‘tree like’ about them. And yet, these connections exist because humanity developed language and systems of communication. Why do these connections exist in this way? What is that connection; the strand between words and images? Perhaps that connection is something intangible; after all, humanity created its lexicon as a way to quickly communicate ideas through signifiers and signified; there is nothing inherently ‘tree like’ about a tree, we just made up that word somewhere along the way. But this connection still exists between the written word and the visual image, between the thought and the idea of a tree, the manifestation of that idea as written language, and the re-encoding of that idea back into visual imagery when we draw the image of a tree. In many ways, this cyclical relationship between ideas, words and images mirrors the rhetorical situation writers so often concern themselves with; the cyclical connection between author, audience and text. So what happens when we use technology to create images based on our words? What happens when we remove the human filter and add an algorithmic one? In what ways does AI interpret, decode, revise, filter, analyze and create images from words? And what does this mean for our understanding of how writing works? I believe, and hope to demonstrate, the crucial role artificial intelligence will play in the reshaping traditional notions of rhetoric, ownership, authorship and the social construction of writing. So join me on a journey through the art AI application Midjourney and let’s spelunk down this rabbit hole together as we think about words and images.
According to IBM,
“Artificial intelligence leverages computers and machines to mimic the problem-solving and decision-making capabilities of the human mind.”
In writing studies, the concept of rhetoric is seen as the dynamic relationship between an author, a text and an audience. As we study rhetoric and its applications for writers, we often focus on the conscious choices writers make as they design texts; the specific moves writers make to achieve certain goals in their compositions. These goals range from accessibility considerations, understanding and adhering to various genres in order to engage with audiences, using technology to create compelling texts and many more. Writing studies students are encouraged to think about their choices, reflect on the rhetorical moves they make in their texts and develop strategies to use writing skills in specific ways through rhetorical frameworks. And here we have AI; computer systems that are attempting to mimic human decision-making capabilities; the very thing writing studies scholars and students are trying to develop. So how is AI going to disrupt our ideas about rhetoric? If computers are creating for us, how will writing change?
Let’s take a quick look at the AI Bot called Midjourney. Midjourney is a very powerful AI art bot capable of creating anything imaginable. The AI learns and grows and gets better as more and more people use it. What fascinates me about Midjourney is the fact that it essentially runs on words and writing. In order to make this bot work, users write. Text prompts are fed into this bot and the AI creates original artwork in seconds. Users can revise, share, edit and tweak their creations by manipulating the words they write to the bot. For example, the prompt /imagine “a robot typing a digital essay” produced this set of images:
So this is a direct connection between signifier and signified; in this case the words “a robot typing a digital essay” was decoded/encoded/interpreted by the AI to produce those images. And if we roll that input again, we get these:
Two very unique sets of images based on the same set of words. The AI mediates this message in the same way writers mediate messages with the rhetorical moves they made as they compose; there are decisions being made, references being…referenced and conclusions being drawn about what this ‘composition’ is going to be. So let’s continue this little experiment and revise our prompt. Just as writers revise their compositions to address audience needs and concerns, writers using AI to make art can do the same thing to more directly influence the output. So let’s change “robot typing a digital essay” to “a robot typing a digital essay under a starry sky”. We get this:
As you can see, the images change and evolve as the writing is tweaked and revised. The idea that words prompted these images is fascinating to me. Without the writing prompt, these images would not exist. On top of that, these images are mediated by a computer that is referencing and sampling data to understand what I wrote to it and then interpret that data to give me what it thinks I want. The connections between this process, rhetoric and writing are crystal clear. But Midjourney, just like writing, can be a socially situated action and this takes the entire discussion to a whole new place.
If you visit the Midjourney Discord server , you will find about 5, 463, 8780 members all creating AI art together. This is a massive community. There are various channels for different images prompts, discussion forums, and support channels. As users create art and ask the AI to create it, all the other users in that particular channel are seeing that art being created in real time; they can see the words and then they can see the images those words produced. What’s more, users can create variations or upscales of someone else's’ creation. This fascinating process of creation, revision, remix and publication encapsulates so much of the writing process in space of a few seconds. As users are inspired by the art of other users, they create their own art, remix something they like, use the words to create something new, and the cycle continues; and endless stream of original AI art all brought to life through writing.
So what do we take away from all this? My initial question focused on how AI understands writing and in turn, how AI creates original art from the way users use language and writing to interact with the program. Based on my own experiences with Midjourney, it would seem writers have a major role to play in the creation of original AI art. Writing is a continue, evolving and process based activity. Learning this process and applying it to the use of AI can make for better and better creations from the bot. Knowing how to think about words rhetorically, not for a human audience but to more effectively communicate with a robotic one, will become more and more important as AI continues to find its way into the mainstream. And the possibilities are as limitless as the art being created in Midjourney. How might writers use this technology to create more interested compositions?
I continue to think about the idea of ‘writing a painting’ and I think this is why Midjourney is so appealing to someone like me. I love writing and I love being creative, superficially with art. AI art combines these two passions into something new, some new kind of creative production. Maybe we don’t have a name for this new thing yet; perhaps the signifier is yet to emerge. But either way, we are in a new era of using words and writing to create beyond the page, beyond the signifier, beyond the confines of alphanumeric language and who knows where this journey will end.
If you have a chance to try the free trial version of Midjourney or other AI art bots, post your creations on the Wall below and thank you for reading this piece.
You can find me via email at jcary1@uncc.edu.
