AI Doesn't Prevent Students From Learning. It Shatters the Illusion That Schools Were Measuring Education.
- Jul 3
- 6 min read
What exactly is learning?
Is it a process or a mechanism? Does it only occur in specific places, like museums and classrooms? Is it fixed, or is it transient?
The definition of what constitutes learning seems to have been eroded from cultural memory and replaced with the idea of course completion or earning a grade.
When I entered college, I was genuinely surprised and confused by how colleges operated … unofficially. There were opportunities to pay for essays, purchase class notes, join a sorority to get copies of past exams, or look up past exams online. Students could simply "C" their way through their degree programs. It was a startling realization, having been raised to believe that education was both a privilege and the "great equalizer."
Now AI makes it easier than ever not only to pass, but also to graduate with distinction. After all, nearly the entire undergraduate curriculum now sits behind a prompt. While Google made the material searchable, students still had to assemble it. Then ChatGPT responded to the prompt with a new offer: “I can retrieve it, assemble it, explain it, and walk you through it.”
How I actually learned
As a child, I learned how to count with an abacus, and I learned to spell at home with Scrabble tiles. Children today might not be using such devices, but they're still given access to the content. One plus one is still two whether I count it on beads, on an app made for preschoolers, or using ChatGPT.
But how did I learn?
First, there was someone there to teach me, guide me, and correct me. Second, I was engaged, because I liked the colors of the beads and I loved putting the Scrabble tiles together. Third, I was asked to repeat what I learned out loud. Count to twenty, Sophie. Or, what words did we make today? The result was that I learned.
All of this can be replicated with or without AI. If ChatGPT guides a user through colored beads or Scrabble tiles, acting as a tutor, learning can occur if there is structure around it: a parent or educator involved, interaction from the student, and a way to measure the learning that isn't a Scantron, a multiple-choice quiz, or a participation grade.
The problem isn’t that AI prevents learning. Rather, the systems claiming to measure learning need new tools.
Who still learns?
Educators have found themselves in a precarious position. Younger children are more exposed than ever to technology, including AI, in the classroom. As a result, the teacher’s responsibility has shifted toward facilitating that technology more than ensuring that learning has actually taken place. With frequent testing used to determine whether students have reached benchmarks, education at the primary and elementary level is becoming more about passing children onward than verifying that they have internalized what they were taught.
Colleges and universities face an even greater challenge, because young adults have more autonomy than younger children. Students today can not only purchase copies of old exams; they can also use ChatGPT to create the very artifacts used to measure their learning. And therein lies the entire problem with the AI-in-education crisis. The essay, the take-home midterm, and the book report can all be solved with a prompt to an AI chatbot.
So in the modern classroom, who, if anyone, still learns?
Schools deliver credentials
Schools are structured around credentialing. It's a hard truth for many to entertain, but it is foundational to understanding why learning itself was never what was measured. Schooling is the institution; education is the deeper human formation; learning is the active acquisition that feeds it. Schools provide credentials to demonstrate basic competence, credentials to graduate, credentials to enter the job market, and, once employed, credentials to advance in one's career.
As a result, our society created scales to measure who performed better on the tests, wrote the best essays, or had the best math brain. These are often referred to as rubrics, instruments that are not nuanced enough to verify learning but are sufficient to rank performance.
And it worked for decades, until an articulate chatbot responded to a prompt with, "ask me anything."
When submission replaced learning
Sometimes society creates its own dilemmas.
As a college student, I completed computer science courses as electives. They weren't relevant to my biology degree, but I had enjoyed programming classes in high school as a teenager. My scholarship absorbed the cost, so I enrolled.
Much to my surprise and dismay, the experience was quite different from my earlier experiences and expectations. In the college classroom, there was no requirement to write the code on paper before gaining access to the computer, and no teacher checking progress as students worked individually. Assignments weren't graded manually with handwritten comments on how my code could be better. There was no high five when I figured out how to do it correctly.
In high school, I had learned both how to write computer programs and how to demonstrate my comprehension of the subject matter. In college, I had witnessed programming education reduced to assignment submission. Students shared their assignments freely among each other and passed easily. The system had trained students to value the output over the process.
What's the dilemma?
We eventually trained a generation to be "output managers" rather than thinkers.
In those college courses, coding became less about understanding the code and more about getting it to work. Consequently, an entire generation of professionals carried that belief into the workforce. They didn't need to understand the code deeply to get hired as coders. They needed to "make it work." Then the online coding repositories filled up with reusable code, and the online coding forums filled with shared answers.
Today's AI has absorbed the online forums, repositories, examples, shortcuts, and shared answers that shaped modern coding practice. It produces the code, while students and professionals alike wonder where that leaves them. And society is left wondering how computer science, once considered a sure career choice, became a questionable career path.
AI automated coding easily because we had already been treating computer programming as output generation instead of, for lack of a better term, "deep learning."
From tests to shadowing
Maybe the fallacy in our approach stems from treating knowledge as something that can be packaged, transferred, and measured apart from practice. Once we bound it into books, knowledge became portable and widely accessible. Students could learn without the expert behind the words being present in the room.
Many programs solved the need for real-world experience with hands-on labs and the practicum, a supervised, hands-on application of study in a real working setting. Perhaps leaning even further into the apprenticeship model will help restore the observation, correction, and guided repetition that learning requires. Want to become an attorney? Spend time shadowing one.
The apprenticeship model might also alleviate the related challenge AI presents: how we enlist junior employees into the workforce now that AI systems are becoming increasingly fluent at performing entry-level tasks. Because AI can perform that work, new entrants to the workforce are increasingly expected to step directly into roles requiring higher-level judgment, critical thinking, and the ability to oversee and manage AI systems. Therefore, apprenticeships with more direct observation and instruction might help junior employees build the judgment and practical experience that entry-level tasks once developed more gradually.
Education is a lifetime experience
AI threatens how we've been schooling students, not the students' ability to learn. It does so because it can produce the artifacts of learning with little effort from the student. But true learning, the engaged, interactive repetition I experienced as a child, cannot be automated away from the human brain. The beads still have to be counted.
Going to Starbucks, trying a new latte with new spices, and updating my beverage preferences is learning. It might not help me land a new job, but it enriches my life. And maybe, as a society, AI is forcing us to rebuild from scratch how we prepare our children for futures we ourselves haven't seen.
Even creating the illustrative art for my articles has been a learning experience. I’ve learned art and photography techniques from AI while using AI to express my own visions. None of this was possible for me before I sat down to learn from and work with AI models like Claude, ChatGPT, and Gemini.
Learning, after all, is simply the active, engaged processing of new information to update our understanding of the world. These learning experiences, accumulated over a lifetime, become our education. They are found in the abacus beads, the Scrabble tiles, the spices in a morning coffee, and apparently in the process of creating AI-generated artwork to illustrate our stories. AI might automate the multiple-choice quizzes and the take-home midterms, but it can never automate the human experience of discovery.
Join the Conversation
"AI is the tool, but the vision is human." — Sophia B.
👉 For weekly insights on navigating our AI-driven world, subscribe to AI & Me:
Let’s Connect
I’m exploring how generative AI is reshaping storytelling, science, and art — especially for those of us outside traditional creative industries.
About the Author
Sophia Banton is an AI leader working at the intersection of AI strategy, communication, and human impact. With a background in bioinformatics, public health, and data science, she brings a grounded, cross-disciplinary perspective to the adoption of emerging technologies.
Beyond technical applications, she explores GenAI’s creative potential through storytelling and short-form video, using experimentation to understand how generative models are reshaping narrative, communication, and visual expression.
