AI Is Frankenstein’s Monster to Traditional Schooling

All across the country, school districts are creating policies regarding artificial intelligence. The big shots in the central offices seem to enjoy making policies.

And I’m not opposed to policies per se. But any policies on the use of AI are going to be mostly pointless unless the systems seek first a classroom-level understanding of what AI is changing about teaching and learning.

To traditional schooling, AI is Frankenstein’s monster. It threatens to tear to shreds the memory-based epistemological practices on which schooling has forever depended. If learning means remembering information and reproducing it in some predictable manner, AI already outperforms just about anything we ask students to do.

Some AI models have demonstrated performance at approximately the 90th percentile on a simulated Uniform Bar Exam and the 93rd percentile on SAT Evidence-Based Reading and Writing. They can write the essay, solve the problem, summarize the chapter, create the presentation, and produce the code.

“Getting the right answer from a chatbot can create the illusion of learning.”

The problem we really have to face is that many assignments in today’s American education system are designed around an impoverished understanding of learning in the first place.

This warning comes from the recent report of MIT’s Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training. The report describes a growing concern that students can use AI to produce credible answers without performing the cognitive work required to understand what those answers mean.

I’m thrilled that MIT has published this committee’s report because I think it’s a worthy read for every K–12 school administrator in the United States. But let’s be clear about something: AI has not created the distance between completing an assignment and understanding something. It has exposed it.

Traditional models of instruction tend to emphasize the doing: the observable skill, the measurable behavior, the predictable product. I’m not opposed to content standards. But content standards tell us what students should be able to do. And traditional assessments ask them to demonstrate that doing in some prescriptive, pre-approved form. We educators then treat the product as evidence that understanding has occurred.

But the observable action is not the same thing as understanding. It is just a piece.

“When AI makes it possible to offload the cognitive work of learning, how can we assess what students actually know and understand?”

This is the question that should drive district-level conversations about AI. Before deciding whether AI should be prohibited, permitted, or required, educators must become much clearer about the human understanding a learning experience is intended to develop.

What do we want students to understand?

I was giddy to see that the MIT committee recommends backward planning, the McTighe and Wiggins approach of beginning with what students should know, be able to do, and value. This is also the essence of the Culturally Responsive Education mental model.

An integrated understanding includes the doing, but it also includes the conceptual knowing and value-based beliefs that enrich and enliven knowledge to the point of understanding.

When a student understands in this way, they don’t merely memorize parts; they recognize the whole. And their understandings are durable enough to travel across contexts. An understanding brings together the head and the heart so that students can interpret, connect, personalize, and use what they have learned to make sense of new circumstances.

“Instructors need to revisit what they really want students to know and devise assessments that foster, or even include, the kind of productive struggle that builds durable understanding.”

Productive struggle! Kudos to this committee.

Because if we’re talking about struggle, that cognitive friction, the disjuncture that signals a learning opportunity, is also what we mean by rigor.

In traditional classrooms, rigor has too often been confused with difficulty. The more information students must memorize, the more demanding the assignment appears, and the more rigorous traditional methods claim it to be.

But I like to say that rigor isn’t a noun. We have to think of it as a verb. It’s something the learner does.

A person who understands something rigorously can convey its meaning in different ways, to different audiences, in different spaces, with different references, illustrations, analogies, and experiences. They can explain a concept to someone who does or does not share the same background knowledge. Because they understand the essence of a thing, they can recognize its relationship to something that might seem unrelated to someone who doesn’t understand.

When someone understands rigorously, they can draw upon their life biography and cultural fluencies to make the concept meaningful without losing its conceptual integrity.

Real understanding doesn’t come in a one-size-fits-all package because human minds don’t make meaning in identical, cookie-cutter ways. Students bring different experiences, identities, associations, languages, interests, and cultural repertoires into the learning environment. Those differences aren’t distractions from rigorous understanding. They’re its primary resources.

“Education is a cultural practice through which students learn to make meaning, exercise judgment, form identities, and participate responsibly in community.”

It’s been a not-too-well-kept secret that some teachers, especially in some of the more affluent schools, have been going the extra mile to prevent students from photographing, photocopying, or taking home paper-based tests that they give year in and year out. Their problem has been preserving an assessment, presumably to save themselves the time and effort of having to recreate it.

The question facing schools now shouldn’t be how to preserve traditional assessments and assignments from AI. We should be talking about how to develop human beings who can navigate a world containing vast reservoirs of information while still exercising judgment, sustaining relationships, acting with agency, and building lives in which they can flourish.

That requires a more human-centered goal for education.

But if content coverage is the goal, be assured that AI can replace just about every assignment we can invent. But if meaning-making is the goal, students’ humanity becomes the primary resource through which understanding is shaped and delivered.

The goal of education in the age of AI must be meaning-making and not merely content coverage. All AI policies should follow from there.