Some of the best teaching advice I’ve heard recently came from a fictional doctor during a streaming binge.
In Episode 6 (Season 1) of The Pitt, Chief Robbie reminds his team to chart everything: “In your medical records, make sure that your decision making and your notes reflect not just the diagnosis but all the thought you put into ruling out all the critical illnesses” (Adarkwa & Marcano, 2025). The camera cuts away, but that line stuck with me. The kind of charting he’s talking about isn’t just for record-keeping. It’s a way of making the doctors’ thinking visible. It’s a practice that reveals process as much as product. And, in literacy instruction, teaching that kind of visible thinking is essential.
As educators, we know that literacy has always been more than decoding text or producing words on a page. It’s about questioning, connecting, and constructing meaning. Now, with generative AI capable of producing instant drafts and summaries, the need to make thinking visible has never been greater. What if we asked students to “chart” their writing process the way ER doctors chart their cases? Not just the final product, but when, where, why, and how they turned to AI throughout? Not to catch them, but to understand their decision making and to help them see themselves as literate thinkers.
Two years ago, I caught a student using AI to write a poem about his favorite sport. Let’s call him Jon. Jon’s rhyme scheme and word choice were unmistakably artificial. When I confronted him after class, he didn’t deny it. In fact, he seemed confused by my concern. “You always tell us to use mentor texts when we write,” Jon said. “So I thought this could be one. I just had to make it my own.”
At the end of this past school year, I had a different kind of encounter. A student, let’s call her Jane, turned in a short story modeled after her favorite independent choice novel. There were no immediate red flags: no artificial phrasing or strange syntax. But during a Question Flood activity (where students talk through their drafts without looking at them), Jane couldn’t answer basic questions about her own writing. When I checked in after class, she insisted she’d written the draft without AI. “I was just really tired,” Jane said. “I kinda forgot what I wrote.” Although I had my suspicions, I couldn’t prove that her writing was AI-generated. But I later learned that proving it wasn’t the point—charting her thinking going forward was.
These encounters, two years apart, taught me that we can’t wait for the system to catch up. As AI evolves, so must our teaching.
The System Isn’t Ready, But We Can Be
I knew I couldn’t get ahead of AI shortcuts with better plagiarism detectors (which are largely unreliable) or so-called “AI-proof” assignments. But one shift that’s helped me navigate this moment is reimagining Bloom’s Taxonomy for the era of AI. Traditionally, “Create” sits at the top of Bloom’s triangle—built on remembering, understanding, applying, analyzing, and evaluating. But now, many students ask AI to “create” a draft before they’ve engaged in deeper thinking.
My goal isn’t to start with AI-generated writing. It’s to keep students composing at the beginning, pen to paper. But when AI-generated text enters the process (whether at the beginning or midway through), I want them to treat it as a draft in need of refinement.
So I’ve repositioned the work: Regardless of where the text originates, students first evaluate what’s been created, then analyze the tone and structure, then apply changes until the piece reflects their deliberate choices, and finally reflect on what this process reveals about their thinking. It’s not a perfect inversion, but it’s helpful. Students need to see that any AI-assisted writing is still just a starting point. The evaluation, analysis, and reflection are the true finish line.
Students need to see that any AI-assisted writing is still just a starting point. The evaluation, analysis, and reflection are the true finish line.
Charting the Process
While my coauthor and I didn’t explicitly frame our book AI in the Writing Workshop: Finding the Write Balance (Peterson & Magliozzi, 2025) around this inverted taxonomy, the approach is embedded in our pedagogy. We leveraged it to develop five rules for using AI in the writing process. These rules protect the higher-order thinking Bloom’s Taxonomy was designed to prioritize, especially when AI makes “instant creation” feel like an easy first step.
Write first. Our ideal sequence is human-first: students composing in notebooks before they even open a chatbot—drafting, wrestling, and generating ideas themselves.
Struggle second. Ideally in class, in their notebooks. During this stage, students might reread a quick write and underline a favorite sentence. They might circle a passage that needs more context or rewrite a lead until it feels right. But also, I teach teenagers, who will sometimes take the path of least resistance. This is where the rules matter most. I can’t control whether a student outsources the cognitive work to AI—and I don’t want a classroom built on surveillance. But what I can do is create a routine where students are expected to leave evidence of thinking at each step. If someone skips the struggle at home and shows up with a polished draft, the process isn’t over.
Prompt third. Question fourth. Reflect and be transparent. These final three rules aren’t a permission slip to always create with AI. They’re a scaffold that emphasizes, “If you choose to use AI somewhere during this process, what it produces is never the end product.” They must evaluate what’s there, identify what’s missing, reshape the voice and structure, add specificity and truth, and document the decisions they make along the way.
Thinking Alongside AI
My students now have unfettered access to Gemini and NotebookLM. To give them more granular guidance, I’ve aligned my reflective rubric with Marilyn Pryle’s “Five Questions” framework (2025). Originally designed for text analysis, Pryle’s questions transfer beautifully to AI-supported writing.
1. “What am I reading?” Students review the output and identify what it is: A brainstorm list for five possible titles? A summary of a chapter? A draft of an essay? They determine the output’s relevance to the task at hand.
2. & 3. “What is it showing?” and “What is it hiding?” Students examine the AI’s arguments, omissions, and possible biases. What’s emphasized? What’s missing? What doesn’t feel right?
4. “How am I reacting?” Students reflect on whether the AI supports or challenges their thinking. Are they excited by the possibilities or suspicious of the output? What’s still unresolved?
5. “How does it work?” Students assess the quality of the response. Instead of copying and pasting, they do the work of making the output stronger: integrating ideas from class discussions, quick writes, and primary texts; revising the structure; reordering for clarity; and ensuring the piece meets the requirements of the assigned task.
By charting these questions more intentionally throughout the writing process instead of only at the end, students develop a stronger vocabulary for interacting with AI as a thought partner they manage rather than a voice they borrow from. Figure 1 (p. 31) shows how Pryle’s Five Questions align with the stages of my reflective rubric for AI-supported writing.
The Real Work Still Belongs to Them
One of the most powerful moves I’ve made is to ask my students to reflect not only on what they wrote, but how and why they wrote it. These reflections happen on a Google doc in class the day the summative is due. Students also make a case for the grade they deserve. While some students choose to answer these questions in short phrases, most come to see that this level of documentation isn’t just busywork, but leverage. It gives them the language and evidence to justify the grade they believe they’ve earned, because they can point to specific choices, revisions, and moments of decision making rather than providing the standard, “I worked really hard on this” response.
Ayla, a 9th grader, reflected on her slam poem “Scars of Innocence,” sharing that she wanted to trace the shift from childhood’s physical wounds to deeper emotional scars. After drafting the entire poem, she turned to AI for one specific purpose: to brainstorm stronger titles. She knew her original title wasn’t quite capturing her theme of transformation, so she used AI as one tool in a much larger creative process—not as a replacement for her own voice, but as a way to refine it.
Then there’s Eliza who described using AI as a brainstorming partner: “I asked what I could do to strengthen and lengthen my poem, and I was very happy with the results that I got from it. Though I did not take any lines directly from the bot, it is what initially gave me the idea of using an extended metaphor in my poem.”
That’s the literacy work. That’s the learning. AI might provide students with a product, but our job is to ensure students still own the process. In a world where creation can happen in seconds, it’s the charting—the reflection, documentation, and transparency—that shows us what they’ve really learned. Some students may push back on documenting their process or try to treat AI output as a final draft. But the point of this framework is to make that move harder to hide and easier to revise—and to keep the most human part of writing, the thinking, at the center. AI isn’t the problem. Uncharted thinking is.
