Higher education needs better understanding of relationship between AI and literacy, article argues

LAWRENCE — As with most companies and enterprises, higher education is facing the same pressure to integrate artificial intelligence into its framework. But in the rush to embrace this innovation, institutions may be promoting technology for technology’s sake rather than truly navigating the positive and negative aspects for reading and writing.

“Universities are in this race to promote their AI literacy offerings and all the various cool ways they are integrating text-generative AI — not just in the classroom but research, admin, anywhere writing is happening. I would prefer universities take a step back and look at the larger picture of literacy’s role in the curriculum,” said Sean Kamperman, assistant professor of English at the University of Kansas. 

His article titled “On ‘AI Literacy’: The Rhetorical and Historical Fallacies of a Movement and Its Marketing” critiques the uncritical adoption of AI education frameworks by exposing the false promises and historical distortions used to sell them. The article appears in Change: The Magazine of Higher Learning.

“The term AI literacy is shorthand for learning the basics of artificial intelligence technologies that are rapidly changing, based on the premise that AI is inevitable and here to stay. Part of what we object to in the article is the way that the term is bandied about without any real conceptual grounding in literacy itself,” said Kamperman, who co-wrote the piece with Harvey Graff of Ohio State University.

Currently, he said universities are scratching the surface of what AI actually means, which is problematic given the broad definition of such tech. 

“Most of the time when people in higher ed refer to AI, they seem to be referring to large language models (LLMs) or chatbots and synthetic text generators,” he said. “Many of the AI literacy programs we looked at teach individuals some basics about how that stuff works, occasionally about limitations and failure modes. Then usually there’s some perfunctory discussion in there of ethical use. But use is nearly always assumed.”

Kamperman and Graff instead encourage a three-step guideline and approach:

• AI literacy initiatives in higher education too often conflate technology skills with reading and writing, obscuring the complex ways AI and literacy interact.
• Historical study of literacy shows there are two myths AI literacy initiatives tend to promote: (a) effective use of today’s generative AI applications is a foundational skill that is simple to acquire; and (b) AI literacy is a sure path to economic security.
• The professors advocate for an “AI and Literacy” approach informed by historical understanding of past writing technology innovations that centers reading and writing in learning about generative AI tools.

“We need to be thinking hard about the affordances and limitations of AI for reading and writing,” Kamperman said.

“For decades, we didn’t have to explain to students or indeed ourselves, ‘Why am I having you write a literary analysis essay in this class?’ or ‘Why am I having you write a scientific paper?’ This is just what we do. Now that there’s a technology that can purportedly do all of these things for us, we need to explain what the point is of doing the thing in the first place.”

He said universities should think about how AI can fit into that picture. How can it enhance or disrupt whatever goals or skills educators are trying to achieve with those particular forms?

“People are using LLMs extensively in their reading and writing to try to scale what they do and become more productive. That might become the norm, and we might discover in five or 10 years that it has tangible social benefits in some areas,” he said. “But sadly we will probably also discover it erodes important literacy skills and increases dependency on technologies controlled by a few powerful corporations.”

A KU professor since 2022, Kamperman specializes in rhetoric, professional and technical writing, digital media studies and disability studies. He is also co-director of the KU AI & Digital Literacy Institute, a professional development program for high school and college humanities instructors that aims to spur dialogue around ethical and use considerations related to artificial intelligence in education.

“In a couple years, we’re going to have the first generation of Ph.D. students who are ChatGPT natives,” he said.

“For instance, a student who is writing a grant proposal or a research prospectus might go to a trusted platform and say, ‘I need to put these ideas in a specific format. It needs to have these sections.’ People are always looking for ways to cut corners, and something like ChatGPT really does oftentimes offer a way to cut corners.”

However, that tech introduces a whole new standard of literacy to the academic environment … for both students and professors.

“If I use AI to write significant portions of an article or a book chapter, how is that rewarded by my university?” Kamperman said. “If you have people who are simply producing more using LLMs, and that’s what the university cares about, then all of a sudden the expectations gradually ratchet up for everybody else. There are too many important conversations to be had to rush blindly in; we need more people engaged in the conversation, not more AI for AI’s sake.”