Opinion: Who Is Doing the Thinking?

Student using a computer
(Solé Gomez via Pexels)

David J. Butler

The debate over artificial intelligence in the classroom is usually framed as a choice: ban it or embrace it. That may be the wrong question.

In September, two high school English teachers made opposing cases in The Free Press. Deb Eaton bans AI from her classroom, treating virtually any student use beyond spell-check as academic dishonesty. Rod Naquin has incorporated AI into his teaching and argues that educators who refuse to explore its possibilities are missing an extraordinary opportunity.

Both make persuasive arguments. Together they expose the more important question: What are schools supposed to teach when machines can increasingly do the work we traditionally assigned students in order to teach them?

Eaton’s strongest argument is that struggle is not an unfortunate byproduct of education. It is education. A student wrestling with a thesis, discarding a bad idea, staring at an awkward paragraph and trying again is developing judgment and the ability to think. Even brainstorming matters. Coming up with ideas, rejecting them and starting over is part of discovering not merely what to write, but what one actually believes. AI can make much of that struggle disappear. That is precisely what makes it so useful — and potentially so damaging.

Naquin is nevertheless right that simply banning AI makes little sense. Used properly, it can provide immediate feedback, challenge assumptions and help students test their reasoning. An October report from the University of Minnesota, drawing on the judgments of 30 education researchers, suggests that AI holds particular promise when students use it to practice and strengthen skills, rather than bypass the effort required to acquire them.

That leads to a distinction that is more useful than ban-or-embrace: Are students using AI to enhance a skill, or to avoid acquiring that skill in the first place? And that raises an uncomfortable question. If AI can write a better essay than a student, why insist that the student write the inferior one?

Because the essay was never really the product. The student was.

We assign essays, equations and research projects not because the world desperately needs their answers, but because producing those answers changes the person doing the work. Struggling with an argument develops reasoning. Finding the right words develops precision. Discovering that an idea doesn’t work develops judgment. The capacities remain long after the assignment is forgotten.

That is why an experienced lawyer using AI to test an argument is fundamentally different from a student asking AI to construct one. The lawyer may be outsourcing execution. The student may be outsourcing formation.

We understand this elsewhere. Children learn arithmetic before becoming dependent on calculators. Musicians practice scales. Athletes shoot thousands of free throws despite technology that can analyze their technique. Mastery requires doing difficult things before tools make them easier.

That suggests a useful principle for AI in education: competence first, augmentation second. But even that requires qualification. Technology really does make some skills less important. Schools shouldn’t preserve every old exercise simply because previous generations endured it. The challenge is deciding which human capacities we refuse to allow AI to make unnecessary.

Reading carefully. Reasoning independently. Distinguishing good evidence from bad. Constructing an argument. Recognizing when an answer doesn’t make sense. Expressing an idea clearly. Those skills become more important, not less, when machines can generate plausible answers almost instantly. How does someone fact-check AI without enough knowledge to suspect it is wrong? How does someone improve an AI-generated argument if he cannot construct one himself?

This is also why generative AI is different from earlier classroom technologies. A calculator calculates. A word processor records words. A search engine retrieves information. Generative AI can increasingly formulate the thesis, analyze the information, anticipate objections and write the paper. It can perform much of the intellectual exercise itself.

Schools therefore need to rethink not only how students learn, but how learning is demonstrated. Educators are experimenting with more in-class writing, oral defenses, handwritten outlines, documented revisions and assignments requiring students to explain their reasoning. But schools should go further. They should identify which intellectual skills students must demonstrate independently before permitting AI to assist them. Otherwise, we risk measuring the sophistication of the technology rather than the development of the student.

The question must shift from “Did AI write this?” to the much more important “Did this student learn anything?” That doesn’t require keeping AI outside the schoolhouse. Today’s students will enter workplaces where intelligent AI use will be an advantage and perhaps a necessity. They need AI literacy: how to interrogate its answers, recognize its weaknesses, challenge its conclusions and use it without surrendering judgment.

There will be no universal dividing line. A tool that helps an advanced student refine an argument may spare a beginner from learning how to construct one. Different ages, subjects and assignments will require different boundaries. But the principle should remain constant: AI should extend a student’s intellectual abilities, not substitute for abilities the student has never developed.

The greatest danger of AI in education, then, isn’t cheating. Schools have always dealt with cheating. It is that we will mistake the ability to obtain an intelligent answer for the ability to produce one.

Education has never merely been about producing answers. It is about producing people capable of reaching them.
AI may become one of the greatest educational tools ever invented. But before we celebrate what it can do for students, we need to decide what we still insist students learn to do for themselves.

David J. Butler is an attorney. He is president of Dvash Consulting, LLC and a member of the ownership group of Mid-Atlantic Media, which owns and publishes Philadelphia Jewish Exponent.

LEAVE A REPLY

Please enter your comment!
Please enter your name here