Show Your Work: Essays, AI, and the Future of Thinking

 

When artificial intelligence first became widely available to students, many people immediately understood the problem. A student could open a chatbot, type in a prompt, and receive an essay in seconds. Teachers panicked. Students got excited. Parents were confused. Schools began looking for detectors, policies, punishments, and workarounds. It looked, at first, like the essay itself had been broken.

But maybe the arrival of AI did not break the essay. Maybe it exposed how poorly we had been explaining what an essay was for.

Most students experience essays as assignments. They are papers to be written, pages to be filled, arguments to be structured, citations to be included, and deadlines to be survived. In that model, the essay is mainly a product. The student produces it, the teacher grades it, and everyone moves on.

But that was never the deepest purpose of the essay.

An essay is not merely a container for information. It is a tool for thinking.

The word essay comes from the French essayer, meaning to try or to attempt. An essay is an attempt to understand something. It is not simply a report of what the writer already knows. It is a way of discovering what the writer thinks. A good essay begins with a question, a tension, a confusion, an observation, or a theory that is not yet fully clear. The writing process gives that unclear thing a place to unfold.

This matters because human beings have limited working memory. We cannot easily hold a complex idea in our minds all at once, especially if that idea contains several moving parts, examples, objections, definitions, and implications. An essay becomes an external thinking system. It allows the mind to put pieces down, move them around, compare them, organize them, test them, and refine them.

In that sense, writing is not just communication. Writing is cognition.

A person writing an essay is using language as a kind of external processor. Just as a vehicle extends the body’s ability to travel, writing extends the mind’s ability to think. A person sitting silently with a difficult question may reach a limit. But once the question is written down, divided, expanded, challenged, and reorganized, the mind can go farther than it could alone.

This is why some people write essays voluntarily. Strange as that may sound to students who have only experienced essays as homework, many people write essays because they want to understand something more deeply. The essay is not punishment. It is exploration.

Historically, deep thinking has rarely happened in perfect isolation. Many great thinkers developed their ideas through correspondence, dialogue, and intellectual exchange. Charles Darwin wrote letters constantly as he developed and defended his ideas. Sigmund Freud and Carl Jung developed, challenged, and eventually strained their relationship through intense correspondence. C.S. Lewis and J.R.R. Tolkien shaped each other’s thinking through conversation and critique. Philosophers, scientists, poets, and political thinkers have often relied on letters, salons, journals, lectures, and conversation to test and refine ideas.

The solitary genius alone in a room is only part of the story. Ideas often grow through exchange.

AI changes the speed and accessibility of that exchange.

Instead of waiting weeks for a letter, a student can now have twenty exchanges in an afternoon. Instead of only thinking alone, a student can ask questions, test a thesis, request counterarguments, explore examples, and clarify confusion in real time. This is new, and it is powerful.

It is also dangerous if misunderstood.

The first and most obvious use of AI is also the weakest educationally. A student receives an assignment, asks AI to write the essay, copies the result, and submits it. This workflow is simple:

Assignment → AI → Essay.

The student may receive a paper, but very little thinking has occurred. The essay has been produced, but the purpose of the essay has been bypassed. The result may be polished. It may even be grammatically correct. But it is often empty in the way people now recognize as AI slop: smooth, generic, competent-looking, and strangely bloodless.

Teachers are not wrong to worry about this. If students use AI only to avoid thinking, then AI becomes a ghostwriter. It replaces the very process the essay was meant to create.

But there is another way to use AI.

Instead of treating AI as a ghostwriter, students can treat it as a thinking partner. Not a real person, of course. AI is not conscious. It does not care, understand, or experience the world the way a human being does. But because it can respond conversationally, it can still serve an important function: it can help a student explore thought.

The better workflow looks something like this:

Question → Conversation → Understanding → Essay.

The student begins not by asking for the final paper, but by talking about the topic. What is confusing? What is interesting? What does the student already think? What does the assignment ask? What seems obvious? What seems wrong? What examples come to mind? What does the student’s own life, class discussion, reading, or curiosity bring to the question?

In this model, the AI does not replace the student’s mind. It gives the student’s mind something to push against.

The student might ask AI to explain the topic, but also to challenge their assumptions. They might propose a thesis and ask what is weak about it. They might ask for examples, counterexamples, analogies, or alternate interpretations. They might admit confusion. They might argue back. They might say, “No, that’s not what I mean,” and then clarify. They might discover that their first idea was too broad, too obvious, or simply wrong.

That process matters.

By the time an essay is finally drafted, the student should understand the topic more deeply than when the conversation began. Ideally, the student has learned something. Not merely gathered information, but changed their understanding.

Only then does it make sense to ask AI for help organizing or drafting the essay.

Even then, the student’s role is not finished. The draft should be reviewed, revised, questioned, and personalized. It should be copied into a document, read later with fresh eyes, and adjusted until the student recognizes it as an honest expression of their own developed understanding.

This is very different from letting AI “write your paper.”

The difference is not whether AI helped produce sentences. The difference is whether the student did the thinking.

This suggests a possible way forward for schools. Instead of relying only on AI detectors, which are imperfect and may punish honest students while missing dishonest ones, teachers could shift toward a “show your work” model.

In math class, students are often required to show the steps that led to an answer. The final answer is not enough because the teacher wants to see whether the student understands the process. Essays could move in a similar direction.

A student might submit three things: the final essay, the AI conversation used to develop it, and a short reflection explaining what changed in their understanding. This would not eliminate cheating entirely, because no system does. But it would shift the focus from “Did AI touch this?” to “Did learning happen?”

That is the real question.

If a student submits a polished essay but cannot show any meaningful process, the teacher has reason to doubt the learning. But if a student submits an essay along with a long, messy, thoughtful conversation full of questions, corrections, revisions, confusion, breakthroughs, and reconsiderations, then the teacher can see something important. The student was thinking.

This approach would also teach students a much more useful lesson about AI. The goal is not to avoid AI forever. That is probably unrealistic. The goal is to learn how to use AI without surrendering the mind to it.

AI is a thinking technology, but the human brain is also a thinking technology. Both have strengths. Both have limits. The human mind brings lived experience, curiosity, values, judgment, emotional context, and personal stakes. AI brings speed, responsiveness, breadth, structure, and endless patience. The best use of AI is not to replace the human mind, but to surround it with tools that amplify its abilities.

A backhoe does not make excavation meaningless. It changes what one person can excavate. A calculator does not make mathematics meaningless. It changes which parts of mathematics require human attention. A word processor does not make writing meaningless. It changes the mechanics of revision. AI, used well, may do something similar for thinking.

But tools only help when we understand what the task is.

If the task is merely to produce a document, then AI can do that very quickly. But if the task is to understand, then AI should be used differently. It should be used to ask better questions, expose weak reasoning, expand perspective, test ideas, organize complexity, and help the student arrive at a clearer view.

The essay, then, does not become obsolete in the age of AI. It may become more important.

Because the real purpose of an essay was never to prove that a student could suffer through five paragraphs without help. The purpose was to help the student think more deeply than they could have thought without the essay.

AI can threaten that purpose.

But it can also serve it.

The difference lies in the process.

A bad use of AI produces a paper without understanding.

A good use of AI produces understanding before the paper.

And in the end, that may be the standard schools should care about most.

Not whether a machine helped with the words.

But whether a human mind became clearer.