The papers in this proposed special session ask us to consider how the ‘typed self’ relates to the unexpressed self in the age of AI mediation. Virginia Woolf wrote in Orlando (1928) that “we write, not with the fingers, but with the whole person. The nerve which controls the pen winds itself about every fibre of our being, threads the heart, pierces the liver.” For all the dramatic changes of the last century, the inescapable truth of Woolf’s statement changed less from 1928 to 2022 than it has in the last three years with the launch of ChatGPT. Earlier studies of the composition process, such as that of Vivian Zamel (1982), have emphasized how writing is the process of discovering meaning, while Vaughan Prain (1996) has explored how writing helps the student discover what they believe and who they are. What will the relationship between individual and written product be if the process of self-fashioning is replaced by generative AI’s Reinforcement Learning from Human Feedback, where it is not the human self but the program that is hewn, carved, and sanded into a simulacrum of self?
This panel features four papers by five scholars across five North American institutions on the subject of AI, pedagogy, and selfhood. In “The Essay-Writing Process and the Typed Self”, Nathan Murray will discuss one of the byproducts of a current study underway by his Pedagogy AI Research (PAIR) Group: a corpus of 30 screen recordings of students writing essays in a controlled lab environment, half of whom were permitted to write with the assistance of ChatGPT. By tracing the composition process of three individual students who chafed against the confines of the experiment in distinct ways, Murray will explore how the technology shapes student agency and self-conception in AI-assisted writing environments. This dataset offers rare, real-time insight into students’ writing processes, revealing how they navigate idea generation, revision, and self-expression—with and without AI assistance.
In “We want your ‘authentic voice’...but not if you ‘sound like a robot’”, Whitney Gegg-Harrison explores the alienating experience of being labelled as artificial. Writing teachers claim to value our students’ “authentic voices” most of all. Yet authentic voices are sometimes read, whether by automatic detection tools or by other people, as “AI-generated”. As a result, some feel compelled to distort their own voices so as not to be misread as non-human. What do we even mean by “authentic voice” if writing produced by a human in their own voice cannot be reliably distinguished from writing produced by generative AI? What does it do to a writers’ sense of self when their writing isn’t read as “human”? Gegg-Harrison articulates the potential harms of our rhetoric about AI-generated text, informed by her experiences advocating for students whose work is incorrectly flagged as “AI-generated”.
In his paper “Live the Questions: Responding to AI with the Exploratory Essay”, Matthew Parfitt explores one potential tool for enabling student self-expression and self-fashioning: the question-driven exploratory essay, which demands deep thinking by requiring students to question every premature conclusion. The exploratory essay encourages experimentation with style and voice—things AI, with its statistical predictability, can’t do well. It fosters a constructive use of AI as a tool for inquiry, a “co-intelligence,” a starting point rather than an endpoint.
Finally, in “Self-Resuscitation: Rescuing Students’ Voices in a Chatbot World”, Sheila Batacharya and Amanda Paxton explore how AI use in assessment design and writing development shapes “one's sense of self as writers, for both academics and students?”. Detailing lessons they have designed to support students in finding their voices, Batacharya and Paxton illustrate how they guide students in cultivating their writing selves rather than ceding them to generated text. Their presentation will end with a reflection on the implications of generative text and writerly identity for questions of linguistic equity.