The day she turned 10 months old, my daughter decided she would take her first steps. As my husband and I watched in awe, her little brain made the needed connections between the idea of movement and actual movement. As she put one little foot after another, we cheered her on. She cooed and exclaimed in surprise, as we recorded her newfound achievement.
This moment of growth was an example of a “lightbulb moment.” She had been observing people walking around her and had finally made the connection between the idea of walking and moving her muscles.
As humans, we experience these learning moments at many points in our lives: when we learn to eat, to walk, to talk, or to read, just to name a few. The beauty of these moments comes in the connections made both with people in community and with previous personal knowledge. We build on what we know with the help of others who know more than we do.
I’ve witnessed many lightbulb moments in my more than two decades spent in the university classroom, moments of Christmas-light size and moments as big as a spotlight. My goal is to lead students to these discoveries, precisely because I have thought about and planned for the end goal. As a classroom community, I hope to lead them to knowledge about a certain subject. However, I know they won’t reach that goal without both the text we are studying and a plan to introduce them to specific ideas.
Sometimes these lightbulb moments of connection can take place in a structured classroom environment, but other times they may not. Learning can also happen in the kitchen or the backyard. Yet throughout human history, learning has been seen as so much more than merely the transfer of information. Learning takes place when someone who knows more than we do can lead us into the knowledge or skill we want to have. Even alone with a book, the reader is still benefitting from the knowledge of the author, who has shared it in written form—not to mention all the knowledge that went into the publishing, production, and marketing of that book.
The arrival of easily accessible generative AI tools has raised many questions about the process and purpose of education. Generative AI education schemes promise the possibility of gaining knowledge in a quick and easy fashion.
Arizona State University’s recent announcement of an “AI learning advisor” offers a similar guarantee. While still in its beta version, the ASU Atomic website boasts that monthly subscribers will “learn only what [they] want.” ASU, the public university behemoth that boasts almost 200,000 students, promises that Atomic will put together custom non-credit learning modules in only five minutes. The user is asked: “What’s your main goal or objective?” The user’s answer then prompts the AI to build “custom, self-paced learning modules using ASU course content.”
Faculty have been dismayed that Atomic is scraping their digital material, including videos, slide decks, and assignments, without their permission. The AI agent seems programmed to look for keywords in material within ASU’s online courses, which it then presents in shortened versions (that may or may not be originally in the context of the topic), stringing together information into AI goop that purports to be learning.
Certainly, the US government, book publishers, and other international agencies are concerned about AI companies’ use of copyrighted materials. But perhaps the bigger concern begins at the starting point—where ASU’s advertising promises you can “learn only what you want.”
When a person wants to learn something, whether a new subject or a new skill, he or she may have some knowledge of what that might entail but often does not. Learning how to bake my grandmother’s pecan pie didn’t come from merely following a recipe but also from watching how she would carefully crack the pecans before mixing them in. She taught me how to look for the part of the shell that would extend between the two halves. If any little pieces of it remained and snuck into the pie, they would be a bitter, hard reminder of the refuse that should be kept out. I may have wanted the result of the pecan pie, but mine wouldn’t have tasted the same without her sharing her knowledge which involved things I didn’t know and wasn’t even aware of.
In his novel The Silver Chair,C. S. Lewis gives us a glimpse of what can happen when a person may want to learn something but doesn’t know what questions to ask. At the beginning, the two main characters, Jill Pole and Eustace Clarence Scrubb, are transported into Narnia to accomplish a mission. Jill and Eustace become separated, and while apart, Aslan instructs Jill on the signs the pair are to look for, and he expects her to teach them to Eustace. But when she is finally reunited with Eustace, he knows nothing about Jill’s encounter with Aslan.
The first step in Aslan’s instructions is that Eustace should speak to an old friend he meets. When Jill tries to communicate that to Eustace, he brushes her off and doesn’t listen. Eustace doesn’t know what to ask of Jill, who is to teach him Aslan’s signs. As a result, Eustace misses the opportunity to meet a friend from a previous visit to Narnia, complicating their quest for the rest of the novel and leading to the grief of a missed reunion.
Atomic’s promise to “learn only what you want” limits the student, just as Eustace limited himself by not trusting Jill’s experience even though it didn’t match his own.
He only knew that Jill had been absent and had arrived at an inconvenient time. He didn’t know to ask her about Aslan’s signs, because he wasn’t aware that she had met Aslan. Essentially, Eustace didn’t know what he didn’t know.
In my class this past semester, “Knights, Dragons, and Wardrobes: The Medieval Imagination in C. S. Lewis and J. R. R. Tolkien,” students signed up to take the course to meet a requirement in my university’s general education curriculum. When they began the class, they didn’t know that as we read these author’s works, my objectives for them were not only to learn literary terms or plot structures but also to see how they might encounter real and lasting ideas in these novels. As a result, several students had lightbulb moments when they applied Bilbo’s travels to a real-life situation of recovering from trauma or examined the hiddenness of God in The Voyage of the Dawn Treader.
In much the same way, when we learn—whether about taking first steps, baking a pecan pie, or following Aslan’s way—we must seek to learn in community, with others who know what we don’t know and can lead us in our learning. This is where generative AI falls short, and perhaps always will. For now, at least, it produces only what we know to ask, not what we don’t.
Katherine Cooper Wyma is an Associate Professor of English at Anderson University in South Carolina. You can find more of her writing at her Substack.
The post AI Only Responds to the Questions We Know to Ask appeared first on Christianity Today.