AI in the writing process
HOW AI FITS IN THE WRITING PROCESS WHEN STUDENTS DO THE THINKINGArtificial intelligence is changing how students write, and how teachers teach writing. How can we strike the right balance between reverting to pen and paper and an anything-goes approach to using it? By treating AI not as a shortcut to polished prose but as a tool for sharpening reasoning. The Organization for Economic Co-operation and Development’s AI Literacy Framework offers a useful guide for exploring new perspectives, refining results, and reflecting on thought processes used to cocreate with AI. With such an approach, students stay in the productive struggle of writing with AI tools with the goal of deepening their thinking rather than outsourcing it. Read on to explore how OER Project has aligned classroom strategies and concrete teacher practices with these competencies.Now, we know the DBQ isn’t going anywhere and there are real pressures on students (and teachers) to develop student writing skills in alignment with standard rubrics and formats. We aren’t here to tell you to burn everything down. OER Project means to offer you a few seeds that might help you cultivate a new methodology or two in your classroom. Also, we want to be clear, we aren’t the ultimate AI experts. We’re a group of interested curriculum creators and curators, deeply invested in making your work easier. Any of these practices will fire up student cognition—rather than prime them to plug into a formula—and they can be used before or after a traditional writing process.During the writing processSparring with AIThink of AI as a sparring partner rather than a ghostwriter. Have students draft an outline of an argument, and then prompt AI to identify counterarguments or potential weaknesses. Through a critical conversation, students can test their logic and evidence—just as historians do when they wrestle with complexity. You can model how to frame prompts that push the AI to question, not confirm, a claim—mirroring the kind of skeptical reading historians practice. The goal is to break away from the default “you’re so smart” nature of the chatbots, and train students out of a copy-and-paste mindset.Metacognitive annotationsTo cultivate transparency and ownership of their thinking, invite students to annotate or mind map their process. Ask them to flag where ideas originated—from primary sources, class discussions, or AI exchanges—and to reflect how each shaped their evolving argument. This exercise makes cognition visible and builds the habit of “reflecting on thought processes,” the OECD suggests. It also helps you see how students are using (or misusing) AI in authentic contexts.
2How AI Fits in the Writing Process When Students Do the ThinkingUsing a writing coachAre your students using Khanmigo? OER Project has created a Khanmigo Writing Coach Guide to help teachers keep AI in its proper role: supporting thinking, not replacing it. The guide shows how to prompt Khanmigo to challenge student claims, surface counterarguments, and ask “historian-like” questions. It also includes classroom routines that help students document their process so teachers can see where AI contributed—and where student reasoning took the lead. Think of it as a set of guardrails that keeps AI functioning as a coach, not a coauthor.Shift the task: Design for human insightIf AI can produce a surface-level summary or rehashed historical argument in seconds, we need to design assignments that demand deeper thinking. That means prompts that require human insight, empathy, or value judgments. Assignments like op-eds, blog posts, research proposals, or policy briefs let students write for real audiences and use evidence to persuade, not just perform.You might try out this litmus test for the value of your writing prompts: Would a human’s answer be more meaningful than what ChatGPT would produce? If not, revise the task. Personalize prompts to bring students’ own perspectives into play—questions like, “If you lived through that time period, which political platform would you support? Who would you seek as allies, and why?”Change what you assessAs writing evolves, so should our assessments of thinking.Oral communicationSome educators are turning to oral exams, discussions, or conferences to probe conceptual understanding in real time. These can reveal reasoning in motion—and give insight into areas of nuance and uncertainty students are grappling with. A conversation can reveal student thinking as it’s forming, and strategic teacher questions can shape students’ historical thinking as they navigate their in-the-moment answers.Assess AI conversations You might also evaluate students’ conversations with AI tools. Princeton’s D. Graham Burnett assigned his students a dialogue with a chatbot about the history of attention, which students then edited down and turned in. He confesses that “reading the results, on my living-room couch, turned out to be the most profound experience of my teaching career.” What moved him wasn’t the machine’s responses but his students’ capacity to wrestle with them, using rigorous thinking and a real depth of understanding. That’s what we should aspire to: students who can spar with the bots and still own their ideas.Collaborate with other teachersNo one should navigate this revolution alone. One of the gifts of humanity is our capacity for collective learning. Share approaches, compare student work, and refine prompts with your colleagues. OER Project’s upcoming professional development session History in the Age of AI: Teaching Critical Thinkersis one place to start. And check out conversations in our Community Forum, which we link to on the AI for Teachers page.