From the Classroom: AI and Art of Productive Struggle
by Carl Haber
People often ask me why I became a teacher. Sure, I have the noble answers that I fall back on: “It’s inspiring to work with young people,” or “I want to have an impact on my community.” But the real reason I went into teaching is because I couldn’t imagine my life without a summer break to reset. I eventually realized that my soul craves the cyclical nature of working in a school because it echoes the Jewish calendar’s call to reflect on what’s been working, what hasn’t, and what I want to do differently.
That cycle became especially important as the English department began rethinking how we assessed student writing. Two years ago, we began moving toward standards-based assessment and developing department-wide proficiency scales. This forced us to identify much more clearly what skills we actually want the students to develop. Rather than treating an essay as a large individual task, we began to see these writing assignments as expressions of skills they’ve been working on over the course of their whole careers at Milken: making an arguable claim, selecting strong evidence, analyzing language, explaining how evidence supports an interpretation, and so on.
That clarity exposed a problem in my classroom. Students needed repeated practice with those skills before a major essay, but I could not realistically give meaningful feedback on every practice paragraph they wrote. Like many English teachers, I had often assigned short analytical paragraphs for homework and graded them mostly for completion. Students were practicing, but they were not necessarily getting the feedback they needed to improve. By the time I gave detailed feedback on a final essay, the opportunity for low-stakes revision and growth had largely passed.
““I have become less interested in asking whether AI belongs in the classroom and more interested in asking what problem I am actually trying to solve when I use it.””
This was where AI began to feel useful to me in a way it had not before. During my Macbeth unit, I built a MagicSchool writing tutor around the proficiency scales we were already using in class. Students wrote nightly practice paragraphs that built toward their final thematic argument essay. I designed the bot to evaluate each paragraph against the same five-point proficiency scale I planned to use for the final essay, prioritizing claim, evidence, diction analysis, explanation, and clarity. The bot then gave the students one concrete revision task based on the biggest weakness in their writing. Rather than rewriting a weak paragraph for them, the bot helped the student identify a targeted weakness in their writing and home in on that sentence as an opportunity for revision. Suddenly, in a quick 5-minute warmup, they were each sharpening a vague claim, adding a second quotation where necessary, or zooming in on one word from Shakespeare’s language to explain what it reveals about Macbeth’s character. My homework assignments began to feel like they were actually doing what they were designed to do: help the students practice.
This experience complicated my earlier resistance to generative AI. Entering my first year at Milken, our department had largely focused on pushing back against the technology. We warned students that it could zap their creativity, and we “tsk-tsked” when they inevitably used ChatGPT to outsource AI-vulnerable assignments. I still share many of those concerns. The more the world reduces space for the inefficiency of being human, the more I believe the English classroom should preserve some of that inefficiency. We should allow the first draft to be ugly or awkward. Students need opportunities to wrestle with an idea before someone, or something, smooths it out for them. As I continue to reflect, I have become less interested in asking whether AI belongs in the classroom and more interested in asking what problem I am actually trying to solve when I use it.
““The question is not simply what the technology can do for me, but whether the way I use it helps me become the educator I want to be.””
Some Jewish day schools blow the shofar during the school year’s opening ceremony, foregrounding the spiritual link between Rosh Hashanah and the beginning of the academic calendar. The shofar calls us to wake up, examine our actions, and begin again with greater intention. That feels like the right posture for me as an educator using AI. I do not want to adopt a tool simply because it is new, efficient, or impressive. I want to begin with the educational problem in front of me, decide what parts of the work students still need to struggle through themselves, and then ask whether AI can help me teach those skills more effectively. The question is not simply what the technology can do for me, but whether the way I use it helps me become the educator I want to be.
Carl Haber is in his third year as an English instructor at Milken Community School. He has previously taught social studies, music, and outdoor education.
