How I draft, edit, and stay human in the loop

For years I believed my advantage was “writing.” Lately I’ve realized the real edge was not keystrokes, it was ideas, structure, and voice. AI has not erased those. If anything, it has made them more important. Rather than pretend we are still in a pen and paper world, I have been trying to model what authentic writing looks like now.

We do not protect writing by banning the tools everyone already has. We protect writing by showing what thoughtful use looks like, and by being transparent about our process.

What I am hearing, especially in humanities

Last week, a high school English teacher stopped me. “I can tell when something has been AI generated,” he said, “but I cannot tell when they have collaborated with it thoughtfully. And I do not know what to do with that.”

He is not alone. Across our humanities departments, teachers are working on the fly, trying to maintain academic integrity while recognizing that the old gatekeeping moves, ban the tool and police the draft, do not hold when every student has ChatGPT in their pocket. The fear is real. Are we farming out the exact skills we are supposed to be teaching?

I do not think the answer is choosing between integrity and innovation. It is redefining what integrity looks like when the tools have changed.

How I actually write

I still start the old fashioned way, an outline, a thesis, a few proof points, and usually one sentence I think could be the closer. From there, I treat AI like a colleague, not a ghostwriter.

• Editing help. I ask for a clarity pass, tighten verbs, fix hedging, and check whether my headings are parallel. Here is what I actually typed for this piece: “Revise for clarity and concision. Keep a conversational, hopeful tone similar to my other blog posts. Offer two options for the opening sentence.” I kept one, rejected the other, and moved on.

• Skeptic check. “What would a fair skeptic say after reading this” It surfaces blind spots before I hit publish.

• Reports and formatting. For formal documents, I use AI to turn tables into charts, crunch numbers, and reshape dense text into something readable.

• Speeches. I keep a base grad speech and add school specific stories and names. AI helps blend those elements while keeping the message consistent.

None of this replaces judgment. I accept or reject every change. If a suggestion dulls my voice, it is out. That is the standard. My judgment stays in control. I also disclose what I did, every time. A short note at the end of a post goes a long way with our community and models the behavior we ask of students.

What I encourage for classrooms and staff rooms

The most helpful shift has been moving from “Do not use AI” to “Show your decisions.”

• Model, then mirror. I demo my messy paragraph, ask AI for a clarity edit, then accept or reject in real time while explaining why. Students should bring their draft, try the same process, and compare choices.

• Assess the thinking. Rubrics weight claims, evidence, organization, and audience impact, not who placed the comma.

• Make the process visible. Version histories in Docs or Word, plus brief process notes that list tools used, prompts asked, and choices made, make learning visible and deter abdication of thinking.

• Cite the workflow. Not to catch people out, but to name steps we can teach.

Guardrails that keep the work honest

• No blank page outsourcing. Start with your outline, thesis, or key points.

• Ask precise questions. “Cut 10 percent without losing meaning. Keep my conversational tone.”

• Verify facts. If AI offers a claim, check it before it lands in public.

• Always disclose. If a tool shaped meaning or form, say how.

Is this just cheating with better branding

I have never believed collaboration was cheating. When I wrote a newspaper column, my dad, a retired English teacher, was my unofficial copy desk. He proofread, edited, and offered suggestions on every draft. The byline was still mine because the ideas, voice, and final choices were mine.

Tricia Buckley, and before her Sharon Pierce and Deb Podurgiel, all staff in West Vancouver Schools, have read every blog post here before they were published and provided feedback.

AI sits in that same category for me, a helper, not a ghostwriter, and always subject to human judgment. What changed with AI is speed, scale, and availability. I can get feedback at 11 p.m., run ten drafts in twenty minutes, and the tool is always on. What did not change is my judgment, my responsibility for choices and my name on the work.

If the goal is proving you can type unaided, then yes, tools muddy the waters. Our goal in schools is thinking for real audiences. We have always used supports, outlines, spellcheckers, style guides, writing partners, rubrics and colleagues. The standard should be integrity and evidence of learning, not tool abstinence.

Equity

AI is a ramp, not a shortcut.

It helps stuck writers get moving, the student staring at a blank page who needs a sentence to react to, or the English language learner who can articulate ideas verbally but struggles with syntax. AI can generate that first sentence, and suddenly the student has something to revise, reject, or build on. For strong writers, it is a way to go deeper, test alternate structures, get a skeptic to read, or polish a conclusion without losing momentum.

The equity move is not banning tools for everyone. It is teaching how to use them responsibly, and ensuring access to good instruction is not the new dividing line. When we teach tool literacy, we level up. When we ban tools students already have, we make the learning invisible.

Prompts that actually help

• Clarity pass: “Revise for clarity and concision. Keep a conversational, hopeful tone. Offer two options for the opening sentence.”

• Skeptic lens: “List the strongest fair minded critiques of this piece and one concrete improvement for each.”

• Structure check: “Are these headings parallel? Tell me how to fix them without changing the ideas.”

• Audience flip: “Rewrite the conclusion as guidance to parents in about 120 words.”

• Report polish: “Turn this table into three plain language insights and a simple chart title. Flag any numbers that look inconsistent.”

What I tell our community

• We are pro-writing and pro-truth. We will use modern tools and we will say when we did.

• We value voice. Your voice should be recognizable across drafts and tools.

• We lead with learning. If a tool helps learning, we will teach it. If it replaces thinking, we will not.

If you want more

Last week I facilitated a Hot Topic discussion, “The Future of Writing in an AI World,” at the Canadian K12 School Leadership Summit on Generative AI. 

North Star

I can spend my time lamenting that writing once felt like my competitive edge, or I can double down on the edge that still matters, clear thinking, vivid stories and the courage to be transparent about how we work. That is the blended human and AI writing world I want to model for students and staff.

The teacher who stopped me in the hallway was right to be uncertain. We are all figuring this out in real time. I would rather figure it out in the open, and model a messy and honest process, than pretend the tools do not exist.

AI transparency note: I drafted this post myself, then used ChatGPT and Claude for a clarity edit and a skeptic read. I accepted some wording suggestions and rejected others to preserve voice. The image at the top of the post was created through a series of prompts using Claude.