Composing with AI

Conclusion

GenAI appears in some ways as another attempt to create ‘perfect information’ through technology that libertarian, neoliberal capitalism craves, releasing humanity from the need to reconcile its vast and disparate lived experiences through anything other than markets. The impulse to have perfect information at instant command is similar impulse as the myth of transience: to find the solutions to our problems in what has come before and treat data modeling as a benevolent dictator with a face that diligently uncovers the way forward that past performance lays out for us. Circulation critique of DIKW explains how public concepts of material knowledge—so-called immutable constants, ‘facts,’ or ‘truth’—are so contentious today in this context: DIKW presupposes that human critical attention (i.e. lived experience) organically intervenes in observations becoming action, and that machines can learn to replicate that critical attention. Criticisms of ‘fake news’ and calls for better information literacy over the past ten years imply this matter is ongoing. If we accept that ‘good’ information derived from data leads to positive impact on the future, it is tenable to argue that ‘bad’ information is an impediment, but DIKW logic does not engage in defining ‘good’ or ‘bad’ information as a rule. Furthermore, when knowledge is formed from observations, and information technology accelerates communication, it is logical to assume that the observations and patterns people use to form their worldview originate increasingly in digital sources, particularly when their design facilitates more circulation (Ridolfo and Devoss 2007). This question is particularly relevant to GenAI, who’s entire received data about the world is the product of digital circulation. Hence, Rhetorical DIKW as part of the writing process can not only help students reexamine their writing processes in the context of heightened digital circulation and AI writing, but also open avenues to connect it to active learning and reposition problem-posing pedagogy by using DIKW to connect writing to shaping life and decisions in addition to documenting the world. Writers, artists, and other workers must have a say in the role GenAI will play in their fields. In the meantime, students should understand that they have more responsibility for work they do using AI since they supply the critical and material elements of the knowledge-process, and they are not absolved from collecting observations and organizing them in the process.

For now, two provisional conclusions emerge. First, rhetorical situation and rhetorical ecology remain fruitful to composition pedagogy with or without ChatGPT, and prompts-as-interaction with ChatGPT arguably make ecological and situational concern more important. Forming a prompt is a kind of rhetorical invention (and a writing challenge in itself), and knowing the situation informs students’ ability to judge ChatGPT’s writing. Pedagogical emphasis on rhetorical situation also frames the labor proposition ChatGPT presents: to get better results, one needs to precisely form prompts, requiring an invention and writing process of its own. Those that do that rhetorical work will get better returns, but it is unclear if that ‘saves’ time for the writer or merely increases the expectations of their capability. Second, assignments that primarily ask for summary, ‘comparison,’ or analysis are, in the abstract, scooped by ChatGPT, since they engage in the kind of past- and present-oriented writing (internalization of complex topic or subjects) aimed at assembling data into information that ChatGPT is built to excel at. To challenge ChatGPT (or writers expediently relying on it), assignments that actively combine past experience with a problem and require future-oriented action will be more pedagogically productive.

Either way, Rhetorically-informed DIKW metalanguage helps teachers and students talk about rhetorical situation and ecology ways that invite comparison between the writing process of students as individuals with their own observational and pattern-making skills and lived experience to draw on and ChatGPT as a (blackbox) amalgam of collected texts: both choose data to form the pattern they will present to someone. Ability to critique information on the basis of wisdom it may generate is a difference-making labor a human will provide if GenAI remains a commercial success. Regardless, concern for futurity is an evergreen “skill” for humans.