When I was ten, my grandmother worried about my future.
“I’m not sure many people are journalists anymore, love,” she would say. “There are not many jobs left in writing—the robots are probably doing it all. Maybe choose something else.”
This notion concerned me at ten, but now, at twenty, doubly so—she was disturbingly on the nose about the robots, too.
Artificial intelligence technologies like ChatGPT (the acronym standing for "generative pre-trained transformer") have seeped into the general consciousness of students and instructors worldwide. Generative AI is comprised of algorithms utilized to create content: from essays and audio stories to simulations and coding, bots can engineer art and dish out answers to almost all questions. So, unless you’re part of the offline population, you’ve heard whisperings of a magical technology that can "do everything for you." But what does that mean? And how does this "fix-all" platform influence writing instruction at the University of Pittsburgh?
Associate Professor Annette Vee, Pitt English director of Composition, is interested in answering these questions, as her specialization lies at the intersection of computation and writing. Having been in the automated writing space for years, she’s been leading conversations about the influence of AI on writing instruction since far before the rollout of ChatGPT. Her book, Coding Literacy: How Computer Programming Is Changing Writing (MIT Press, 2017), considers how people write for computers. Now, she explores how computers write for people.
The AI Rollout: Breakneck Speed
Generative AI for writing “has made huge leaps over the past five years,” Vee says. “But the introduction of ChatGPT in November of 2022 was a watershed moment—regular people could operate the interface using natural language rather than programming.” So when research developer OpenAI released this chatbot to the public for testing, it spread like wildfire. “It’s been very, very fast,” Vee says.
In December and April of 2022, Vee shared a survey with students in English Composition courses, assessing their knowledge about platforms like ChatGPT. Student knowledge was extremely limited in the winter, but come April, most students knew about this AI technology.
With instructors creating their syllabi in December and January, the course of correction has been after the fact. “For the most part, professors didn't have policies put in proactively about it,” Vee says. “Much of this work is retroactively responding to students' integration of these technologies in their writing processes.”
Pitt’s Swanson School of Engineering has just approved new academic integrity codes for artificial intelligence legislation. The way the regulations are written has always implied that there are penalties passing off others’ writing as your own, but this is the first time in history that said “other” is not human.
Student Use of AI
So, how are students integrating ChatGPT into their schooling experience? The server is utilized most in a basic question-and-answer format—this spans math equations and history prompts, not limited to any subject. The tool’s use is further seen in projects, discussion board posts, and essay writing.
“Some students are using these systems in ways that are probably not supporting their learning or the learning objectives of the class,” Vee says. “It’s hard because some things are in the space of academic integrity violations.”
Teaching Professor Dana Och, director of undergraduate studies for the Film and Media Studies program, employs what she calls “scaffolding” to combat harmful AI use, layering course concepts throughout the semester and ensuring that student work remains their own. “AI is not good at thinking across assignments,” Och shares.