Through forward-thinking scholarly research published in a prominent journal, a Slippery Rock University faculty member is analyzing the risks of using generative AI for learning and how faculty and students can use AI to its full potential. John Silvestro, an associate professor of languages, literatures, cultures, and writing, was published in the Composition Studies academic journal for his contributing article titled “Generative Artificial Memories: Teaching AI Text Generators as Rhetorical Memory Devices.”
After witnessing the rising skepticism of AI and the potential disruptions to how students retain information and faculty assess learning, Silvestro sought to re-evaluate the use of AI in current educational practices.
“This project came out of things that were happening on campus around 2023,” Silvestro said. “The concerns about AI and what it means for teaching, learning and writing. A part of it for me was that I was not panicking; I was not concerned. I realized that the issue was with how people were framing the technology.”
Memory, and its impact on individual human development and decision making, was a major component that Silvestro wanted to analyze through the lens of AI.
“I have very long been interested in the role that memory plays in how we write,” he said. “I see this all the time with students. You write Post-It notes or take pictures of things on the board, and you use them. And that is memory work. You are storing information so that you can use it in the future.”
Silvestro was exploring this concept before generative AI moved into the conversation. It was about looking at small details, like how a person can leave an item next to their door so that they remember to grab it. He continued to describe such an action as memory work, or what he calls a memory-based decision.
“Eventually, I brought AI and memory work together,” Silvestro said. “We can think about AI as a memory device. It is something that has an insane amount of information dumped into it, and then it regenerates that information to respond to our prompts.”
If AI can serve as a memory device, how can students use it to its full potential without sacrificing individual critical thinking?
Silvestro called back to the similarities between a simple Post-It note and AI, later describing how AI can be used as a foundational tool, never the end-all solution. His research details how AI can act as a support system that can potentially work to enhance a student’s original learning. But, due to the overwhelming size of information that can be regurgitated through generative AI, there is room for error.
“AI is just an artificial memory device for us that other people control,” Silvestro said. “It’s also a lot of writing from the internet, so I think you should think about both of those things. That’s not to say that it’s inherently good or inherently bad. We should just ask questions. I would trust my own notes more than I would trust a spellchecker, and far more than I would trust ChatGPT.”
According to Silvestro, AI is not always a straightforward solution to a user’s problem, so it is important to exercise caution and fact check when needed. But, if used as a support tool, AI can be applicable to many.
“Generative AI is storing information, regenerating it for us, and then offering us support,” Silvestro said. “If you’re going to use AI, you always want to say, ‘I’m trying to support this specific thing.’ Just going to AI and thinking that it’s going to give you what you want is a mistake.”
Silvestro’s research works to promote generative AI in a way that is both supportive and educational for current students, highlighting the importance of using logical analysis when interacting with AI.