TextGenEd: Teaching with Text Generation Technologies
If you’re in search of ideas to help you teach responsible AI use in student writing, or simply to help you frame classroom discussions around AI, you may want to have a look at TextGenEd: Teaching with Text Generation Technologies, an open access, peer-reviewed collection of undergraduate assignments, each introduced by a short essay written by the instructor who designed and implemented it. The collection is part of the WAC Clearinghouse, an “open-access publishing collaborative that draws on the contributions of more than 200 scholars from six continents.”
The editors’ introduction provides a concise, useful history of automated and computational writing from the seventeeth-century cryptographic experiments of Leibniz to present-day large language models and generative pre-trained transformers (GPTs). It also includes thoughtful reflection on the situation of writing instruction in an era when algorithmically driven assistive tools, prone by their very nature to frustrate many time-tested classroom strategies and expectations, are nevertheless increasingly available (if not inescapable, as software applications and platforms such as word processors and search engines embed them in their interfaces), and increasingly likely to be an expected part of students’ writing toolkit in the workplace. Among their most important points is this: “One variable that obscures the future of writing for [writing instructors] is our affinity for writing.” The same could be said for academics as a whole, who at some level typically like to write even if they find writing in general, or any particular writing task, a struggle, because in academia writing is a form of thinking and, crucially, writing-as-thinking is the kind of writing academics get to do. By contrast,
Most people who write do so with limited time, skill, or interest. Writing is stressful and is often done under duress, in high-pressure educational and workplace settings. Automation often promises to deliver us from drudgery and disadvantage and yet rarely delivers. But perhaps automating some aspects of writing will free some writers to choose other forms of expression more inspiring to or effective for them.
We need to be mindful of our investment in writing as we try to determine which parts of the writing process we might yield to AI—and to what extent we have a choice in the matter. Which parts of the writing process can we cede to AI while retaining what we value about writing? We will soon learn if it is tenable to allow students to use AI for some parts of the writing process (e.g., brainstorming and grammar/style checkers) but not for others (e.g., text generation). We may want to embed constraints in our assignments so as not to offload too much of students’ cognitive work to AI. The open question is whether or not these constraints will be possible as AI language models are increasingly integrated into standard writing workflows, or whether students, employers, readers or writers will care about the human or AI origin of prose.
(In a similar vein, the computer scientist Cal Newport points out in a recent New Yorker article that “Writing is hard,” that its complex cognitive demands may help explain the many quirky rituals that some professional writers develop to help them get down to the task and stay focused, and that chatbots such as ChatGPT may work for some writers as “a brain hack used to make the act of writing feel less difficult.” In other words, there may be a legitimate place for such tools, beyond pedestrian functions such as spell-checking, even when students are writing to think.)
The essays and associated assignments in TextGenEd: Teaching with Text Generation Technologies are grouped under five headings: AI Literacy, Creative Explorations, Ethical Considerations, Professional Writing, and Rhetorical Engagements.
The entire collection is available under a Creative Commons Attribution-Noncommercial (CC-BY-NC) license.
