There is a commercial for Gemini—Google’s AI assistant—that I like to show my students at the beginning of the year (Google, 2024). It features a father whose young daughter wants to write a letter to Olympic athlete Sydney McLaughlin-Levrone. “I’m pretty good with words,” says the father, “but this has to be just right.” He then proceeds to use generative AI to write a fan letter on his daughter’s behalf.
When I ask students for their reactions, they are quick to point out that McLaughlin-Levrone would likely have much preferred an imperfect letter—complete with misspellings and grammatical errors, but heartfelt all the same—to something written by a machine. Some of the more astute students revisit the father’s assertion that “this has to be just right.” Where did he get the idea that the letter had to be “just right”? So much so that he resorted to using generative AI?
It could be his competitive nature, the commercial implies. He was a runner and is coaching his elementary-school-age daughter to run competitively, evidently with the goal of beating McLaughlin-Levrone’s world record. But more significant, perhaps, is his feeling that he and his daughter are both inadequate at writing. His being “pretty good with words,” he senses, isn’t enough to help his daughter write to a world-class athlete.
Like the father in the commercial, many students resort to using generative AI because they think they aren’t up to the task at hand. In many cases, students do write to the task at hand but feel their writing isn’t good enough, so they use AI to “improve” it.
The trouble with relying on generative AI is that the writing it produces doesn’t represent students’ authentic voices. How can we build more confidence in our student writers, so they feel that their writing, imperfect as it is, is still superior to “just right”?
Modeling the Process
“Most students,” writes Donald Murray (2004), “have never seen writing being made. They believe that teachers and writers know a magic rite that places words on the page in an order that is full of grace and meaning the first time.” Because students do not know the “magic rite,” they may resort to their own: generative AI.
One way we can disabuse students of the existence of a quick and easy “magic rite” is to model the writing process for them. Students typically examine final products in school, but they don’t get to see the process that led to that product. A teacher can demonstrate the process by projecting a document onto a screen and writing in front of the class, modeling the act of composition. As Alfie Kohn (2004) explains,
It’s one thing to analyze the techniques of a story or an essay, a finished product, but it’s something else again to observe the process of writing. Particularly if the teacher/writer is narrating, explaining the rationale for choosing this word or that sentence structure, students can witness the false starts, the way errors are made and corrected. (p. 184)
We can also model the writing process in the act of drafting. I like to bring in a draft of a piece of writing I’m working on and have the class give me feedback. Inevitably, they see ways to improve my writing, and they observe my own uncertainty about specific sections and words. When I take their feedback seriously, I am taking them, as writers, seriously.
Students learn two lessons here, the first being that even veteran writers struggle with their writing. The second is that if a veteran writer is willing to accept their feedback, then they must be good writers themselves. It’s a confidence booster either way. At the end of class, students will often remark that they had never critiqued a teacher’s writing before. Penny Kittle (2008) writes extensively about this mutual benefit, and my own classroom experience bears it out.
I like to bring in a draft of a piece of writing I’m working on and have the class give me feedback. Inevitably, they see ways to improve my writing. . . . When I take their feedback seriously, I am taking them, as writers, seriously.
Another way to combat the “just right” mentality is to show students that even pieces writers consider complete are not “just right.” I show students rejection emails I have received in response to articles I have submitted for publication. The process behind one article is especially edifying: the first rejection email is terse, but it provides some helpful feedback. The second rejection, after heavy revision on my part, is longer, with more feedback. The final email is a conditional approval—but even when the article was accepted, the editors still wanted further revision. When students look closely at the dates on those emails, they realize they were written over the span of three years.
Feedback Builds Writers
In On Writing, the novelist Stephen King (2010) recalls impaling his rejection letters on a nail above his writing desk. When the nail could no longer support the weight of the letters, he replaced it with a spike. As King continued to write, the form rejection letters began to include more constructive feedback, which he clearly took to heart. The lesson here is that the real “magic rite” students are looking for to create a finished piece is a lengthy cycle of writing, feedback, and revision.
Specific, thoughtful, honest feedback is necessary to building student writers’ confidence. I try to provide what is commonly known as glow-and-grow feedback, which involves both praise and something to improve.
How Praise Works
Teachers who are overwhelmed with student writing often jump right into what needs fixing (the “grow” part of feedback), overlooking the “glow”. But students need to be told what they are doing well. Lynn Holaday (1997) explains that praise helps the unconfident writers in her college course recognize abilities they never knew they had:
A comment about a vivid phrase says it’s possible. Asking a student to tell you more about his or her ideas because you think they are original and interesting says it’s possible. Pointing out how students might help others in areas where they have competence says it’s possible. And all students are good at something—at humor, at rhythm, at metaphor, at word choice. I have never received a paper that could not be praised in some respect. (p. 41)
As Holaday suggests, and as Daniel Pink (2009) corroborates, praise must be specific, sincere, and private to be effective. I find moments to share praise by standing outside my classroom door before the period starts, so I can pull a student aside and quietly tell them what I liked about their writing instead of singling them out in front of the class.
Criticism Promotes Improvement
While praise is important for building writers’ confidence, showing writers how they can improve can also motivate them to keep working. Rather than slowing them down, specific, constructive, critical feedback suggests that the student writer is good enough to be taken seriously.
On long papers, I limit myself to no more than three aspects of the writing I’d like to see improved. (More than that, things can get overwhelming. Remember, we are looking for better, not perfect.) It’s important that the elements I critique are in the student’s power to change. What I will ask of one student may differ from what I ask of another, but I lean more toward the conceptual than the mechanical. That is, if I could wave a wand and change only three things about a student’s writing, it’s unlikely that one of those things would be that all the commas would go in their correct places. Concepts like voice, focus, and complexity are more important and interesting aspects that are more likely to engage students as writers and less likely to be farmed out to AI.
Grading for Growth and Risk
An oft-replicated finding about traditional grading is that it reduces students’ willingness to take risks (Kohn, 2013). Feeling the pressure of wanting a good grade, students who have little confidence in their own writing may succumb to the easy allure of generative AI. While we may not be able to remove grades altogether, we can design grading systems that reward growth and risk.
A pass-fail system for assignments doesn’t lower standards or mean students will do the bare minimum to “pass.” We can make “pass” as high a bar as we see fit.
One system I have experimented with in my classroom is portfolio-based grading. At the end of the quarter, students submitted a portfolio of their writing, along with a written narrative citing specific pieces. Using a clear rubric, they argued for what grade they believed they deserved for the quarter (Abeles & Rubenstein, 2015). To earn an A, students had to demonstrate significant growth or risk-taking, which received more focus than baseline skill, talent, or “perfection.”
I’m currently working with a pass-fail grading system. While my school district doesn’t allow me to teach a pass-fail course, there are no regulations that prohibit me from considering each individual assignment as pass-fail. Students report that the pass-fail system allows them to take more creative risks with their writing.
Lest you think that a pass-fail system lowers standards because students will do the bare minimum to “pass,” remember that we can make “pass” as high a bar as we see fit. Highly selective colleges like Swarthmore have their students take their first semester as pass-fail (Swarthmore, 2026), with the express purpose of encouraging risk-taking. That’s good enough for me.
Good Is Good Enough
If you’re curious, no AI was used in the production of this manuscript—not in the planning, research, drafting, or even editing. You should see the desk on which I am typing this—it’s a mess of books, journals, sticky notes, Coke and seltzer cans, and even the legal pad where I sketched out an outline. This isn’t the best article that I’ve written, but it’s mine. I wrote it. I’m OK with that.
Besides being a writer and a high school English teacher, I’m also the parent of twin 11th-grade students. I don’t think either even knows how to use AI. As a parent, I’ve tried to create an environment that promotes effort and authenticity over the quest for perfection.
One of them is a straight-A student. They write opinion articles for the school newspaper, which I read religiously. Even as I write this, they are upstairs grinding away at a memoir assignment, which they’ve been working on for most of this weekend, for English class. As much as I love to see straight A’s, I’m especially proud that the writing they produce is their own.
My other child has historically struggled with writing. But I’ve seen them grow in confidence, from broken pencils and tears in elementary school, to dictating what I should type for them in middle school, to having me sit beside them as they write in high school. I joke that what they really need is a bobblehead version of me sitting next to them, continually nodding “yes.” But that affirmation means all the difference. Tonight I will sit beside them as they write for history class. “Yes, yes,” I’ll say, “That’s good.” And sometimes good is good enough. It sure beats perfect.
