Joe Essid, University of Richmond
Cady Cummins, University of Richmond
Abstract
Large language models continue to evolve at a far faster pace than policies at colleges and universities. Writing instruction and peer-tutoring, in consequence, will have to change faster still. In six months of testing by the researchers, ChatGPT began to produce prose with ever greater clarity, analysis, and varied (if often formulaic) stylistic choices. At the same time, all AIs tested struggled with copyrighted materials, sometimes refusing to employ them or quoting sources while claiming not to have done so. The authors include preliminary suggestions for those who staff and direct writing centers, specifically methods for adopting generative AI rather than flatly opposing it. We draw from student responses to a campus survey administered in 2023 and 2024, plus one partnership between AI and sixteen first-year students. Such adaptation to AI may prove particularly useful for those helping writers otherwise marginalized by socioeconomic background, neurodiversity, or personal identity. Finally, we advocate getting ahead of any administrative efforts to dictate terms for use of AI that may lead to reduced status, or outright elimination, of human tutors.
Keywords: Generative AI, LLMs, pedagogy, prompt-engineering, praxis, drafts, working conditions, neoliberalism, employment
The AI Invasion of 2022
When OpenAI’s large language model (henceforth, LLM) ChatGPT debuted in late 2022, it generated opinion pieces decrying the end of many things, including the college essay (Marche, 2022). More recent Jeremiads claim that generative AI (henceforth, AI) has the potential to wreck our current model of higher education (Kirschenbaum & Raley, 2024), even “giving up on education, not advancing it” (Warner, 2024).
At times, these prognosticators cited misunderstood events about Meta chatbots being hastily unplugged after the AIs invented a private language. In reality, the bots’ output had simply degenerated into gibberish (Kucera, 2017). Other commenters invoked Hal 9000, the charmingly wicked AI from the 1968 film and novel 2001: A Space Odyssey. In debates on social media, images of Dr. David Bowman flooded feeds, as Hal refused Dave entry to his own spacecraft. In an echo of Melville’s Bartleby, Hal flatly informed the academic/astronaut, “I’m sorry Dave, I’m afraid I can’t do that.”