As Generative AI Speeds Up, I Stay Slow

This essay first appeared in The Important Work, a space for conversations about teaching writing in the era of generative AI.

I return often to advice I received thirty-five years ago, during a summer of student teaching. After observing me in front of my first class, my mentor had one lesson: Trust silence.

He’d witnessed my rookie tendency to fill the space after any question I posed received no immediate response.

“Don’t rush in. Slow down.”

In time, with practice and growth, I felt the calm confidence that his lesson was meant to teach me, a presence more practical than any preparation.

These days I work as a writing coach, and as generative AI speeds up, I stay slow. The project I work on that most highlights the effectiveness of calm presence and patience is the college-application essay. By now, through one-on-one sessions and workshops, I’m sure I’ve helped well over a thousand high school juniors confront this project.

Before I meet them, most students have not had much practice at writing for self-discovery. Writing has been mostly academic, performed mainly in English classes. It has more often been about how, say, Hawthorne uses ambiguity in a chapter of The Scarlet Letter than about recognizing and navigating ambiguities in one’s own life. It has been about outside research, not inside research. Then, as senior year approaches, this consequential personal essay comes into view, a problem for which students have had relatively little preparation.

Procrastination is common. Uncertainty hovers. Self-doubt or just the lack of orientation makes it hard to see that this project can be a beautifully engaging, instructive, and empowering undertaking. This is not “essay” as students have known it. This is an opportunity for storytelling and reflection, with a real audience, and a chance to learn fundamental writing lessons in a genuinely meaningful context—lessons about the power of specificity, about structure, revision, concision, and more. A student can emerge from this endeavor not only more skillful and fluent as a writer, but with resonating self-recognition, and with an appreciation for what writing is and can be for them.

Essential to this experience of writing is student ownership of the project. Before the emergence of generative AI, the main challenges to a student’s ownership were the influential voices of other human beings—the occasional parent or consultant whose own worries about the competition to “get in” informed strategizing that threatened to dictate the student’s focus and direction. What the student “should” write precluded what the student might discover if they inhabited an open space for exploration. But generative AI is not occasional and is not an outside voice. It has infiltrated the writing space. I am writing these words in Google Docs, and if my cursor remains unmoved for four blinks on a new line, Write with Gemini appears in a soft gray font, alongside the keystroke for activation. As I write, a way to bypass any moment of uncertainty about what comes next, and how, is at my fingertips. The accessibility is of course by design: the seeming solution has been placed as close as possible to the problem. It’s not hard to imagine how tempting AI’s invitation is for the seventeen-year-old for whom writing about the self is new. It’s never been easier to become a follower instead of a leader in finding and telling your own story.

At first, I paid little attention to AI. Perhaps this was self-protective, but it didn’t really have a place in the one-on-one relationship-building at the core of my work. The process was too personal, the teaching about authenticity too central. And the emphasis that my colleagues and I put on student ownership, alongside our explicit policies about integrity and honesty, which are aligned with the fraud policies of admissions offices and the Common App, kept the space secure. Pure. That continues to be the case. The trust we build in the spirit of discovery preserves the creative environment.

Yet I’ve recognized my responsibility to learn, if only to meet the question that has been approaching me: what do I, as a writing coach, provide that generative AI can’t?

At root, my work is to help a writer grow. I and my fellow coaches do this through instructive companionship in the development of a meaningful piece of writing. Nowhere in the evolution of a piece is my humanness more significant than at the very start, when a writer—who may not at all feel like a writer—has little or no motivating vision. How to proceed with belief is unclear.

Here, in this space of not-knowing, is where a writer can achieve and feel the most lasting growth. I am not talking about building writing skills—not yet!— but about establishing self-belief. And here is precisely where generative AI beckons, ready to help, eager to offer the way past this emotional standstill. It, too, offers instructive companionship—and fast! A student can get plentiful, encouraging guidance from ChatGPT or Claude much faster than I can give it, and at any time of the day or night.

I’ve put myself in students’ shoes.

Me: Can you help me come up with a good topic for my college app essay?

ChatGPT: Absolutely—this is one of the most important (and hardest) parts of the process, but we can make it manageable.

Our relationship (“we”) begins. ChatGPT serves up a sound description of the goal of this project and numbered steps to get started finding material that is “real and specific” to me. The ideas and exercises for brainstorming are clear and helpful. Then comes an offer: If you want, tell me 2–3 rough ideas (even messy ones), and I’ll help you.

I experiment by responding with both weak, vague ideas and better ones. Faster than I could or would ever deliver to students, ChatGPT serves up guidance, clarifying comparisions, different ways of considering experiences, possible angles (“Pick what feels true”), further exercises, lessons on structure—all with buoying enthusiasm.

It doesn’t take long to see the pattern. ChatGPT consistently provides a demonstration—an “Example Direction,” for instance, which is followed by a justifying parenthetical: “(Just to Show You What Works).”

And always, at the end, comes an offer: If you want, tell me a specific memory (even if it feels small) and I’ll help you turn it into a strong essay idea—or even draft your opening paragraph with you.

The offers keep coming, always prefaced by “If you want.” If I want, I can get help crafting “a polished opening paragraph” or “a full outline tailored to your story.”

How slippery this slope! Whose essay will this be? At some point I will cross the line into committing fraud, according to the policies of admissions offices and of the Common App.

But a different problem concerns me: this need for speed. The AI-powered writing environment is characterized by immediate movement out of spaces of doubt or uncertainty. With its demonstrations and offers, AI is a salesman for progress. I have exactly what you’re looking for! Come right this way!

I know that I am communicating with a predictive, pattern-finding machine, but I can slide into believing otherwise because of the seeming breathless anxiety operating under the enthusiasm.

Meanwhile, in comparison, as a coach, I am purposefully slow. My work, with teenagers especially, has taught me the power of staying put. The value of my humanness is in the patience I bring and I teach, aware that not-knowing-how makes originality more likely. There isn’t a single student I have supported who didn’t benefit greatly from slowing down—who didn’t need to slow down in order to go forward authentically.

At the start, I teach students how to write a set of discrete sentences from their lived experiences. To write minimally but usefully. I’m seeding the most revelatory environment: live, human conversation, a space in which I can hit pause at any moment.

I use the word wait a lot.

“Wait—say that again.”

“Wait—what do you mean by that?”

There is no blinking cursor. The student says the words again, or finds other words to explain what they meant.

Eventually, always, there is an opening, a fresh curiosity about a phrase or choice of words that is distinct.

“Okay, wait,” I say. “Can you now tell me exactly what you see when you picture that moment?”

I call it a traveling act, how an image—what a student sees in their memory from an encounter—appears on the screen in my brain when their words get precise enough. They are not “writing” as they’ve thought of writing. They are unaware of how inventive conversation is, even though they have been practicing this for years—describing to friends exactly what happened, what they thought when it happened, how they felt, what they think about it now.

The distance between us closes as I convey my interest, as I use my experience and imagination as a human being to inhabit their experience in order to understand it.

But what’s happening here, simultaneously, is twofold. They are beginning to feel interesting as they are closing a different distance: between their present self and the details of that past experience, which only they know. They are discovering—taking off the cover and looking in. As one student reflected years ago, “Once I had said it, I was able to see it.” They are hearing themselves as they become what they need to be for this project: students of inner experience.

My job is to be there, with all my humanness, curiosity, and patience, for the micro-movements out of not-knowing and toward wonder, and ultimately to what we call voice. A writer emerges when they find their own words to describe their own experience distinctly.

In his essay “Not Knowing,” Donald Barthelme celebrates the “combinatorial agility of words, the exponential generation of meaning once they’re allowed to go to bed together.” This is the magical technology! My students are not (yet) writing much down, but they are working with these tools, with words, powerfully and efficiently. They are on the way to realizing, as Barthelme puts it, “how much Being we haven’t yet encountered.” Great numbers of students come to the college-app essay project feeling that they have nothing to write about. But the project can teach them otherwise, showing them how much of Being there is in them. And the learning from this recognition is portable, influencing how they meet future projects. The writer grows alongside the writing.

AI doesn’t listen to you with curiosity. AI doesn’t go for a walk the next morning and think about different remarks you made and wonder whether there’s a connection. AI doesn’t wonder. In a chat, AI doesn’t remember that, a few minutes ago, you said something with a different nuance, something similar but with a new slant and tone of voice. AI didn’t see you when you paused and smiled slightly as your eyes drifted up to the left, when you blushed a little in meeting a new thought or memory. AI didn’t feel that shift and didn’t say, “Wait a minute. Hold on. You were just thinking something. What was it?”

Image © claudiodivizia via Canva.com

Allan is Hillside’s founder and a coach.

LEARN MORE ABOUT ALLAN