What a 30-Day Break From AI Taught Me About My Teaching

On a crowded train home after a long day of teaching, I noticed a passenger frantically trying to retrieve an AirPod she’d dropped between the seats. 

My first instinct? Ask ChatGPT for a solution. 

This technological reflex took me by surprise. Why would I even think of ChatGPT in this situation? How could it possibly help? And should I be worried?

This moment sparked “No AI December” (NOAID), an experimental monthlong detox from generative AI tools. Created by my friend Sandro Rybarik and me, what started as a playful experiment turned into a revealing journey into the cognitive costs of convenience—one that I, an advocate for responsible use of AI in education, now urge others to consider.

Why I Embraced AI—and Why I Decided to Take a Break

Since ChatGPT’s launch in November 2022, educators have been divided on whether to embrace or ban AI use in schools. As a writing instructor at the University of New South Wales, I found ChatGPT increasingly useful in my work: inspiring lesson ideas, succinctly summarizing meetings, and speeding up my day-to-day processes. Many of my students found value in generative AI as well—using it to clarify challenging concepts, summarize complex academic reading, and provide feedback on their writing.

But one day a student told me, “I can never seem to recall the research ChatGPT summarizes for me.” And when another admitted using ChatGPT to make their essays “sound more academic,” alarm bells rang. By using these tools to strip away their own voice, my student had bypassed the crucial learning opportunities that revision provides—a chance to tighten their logic, reevaluate evidence, and refine their ideas. While I understand why my students might reach for ChatGPT when writing and revising, I couldn’t help but think of a key principle from one of my favorite books on learning, Make It Stick: “Learning that’s easy is like writing in sand: here today and gone tomorrow.”

The research on the impacts of cognitive offloading with AI tools is still in its early stages but growing. Studies like this one from Microsoft are beginning to raise red flags about AI’s effects on our mental faculties: Though generative AI did improve worker efficiency, this study suggests it may also diminish critical thinking, promoting long-term dependence and weakening independent problem-solving skills. While more evidence is needed, my monthlong AI detox experience seems to align with the current research. I do believe AI offers remarkable potential in education, but its casual overuse could be quietly eroding the valuable mental processes that have come to shape human learning.

What a Month Without AI Taught Me

After I’d used generative AI tools almost daily for the past two years, life without them proved difficult. On the first day, I found myself missing the ease of having AI summarize a journal article or clarify an idea. Almost every bit of technology I used seemed to have AI baked in: Microsoft’s Office Suite, Outlook, Google Docs, LinkedIn, even my phone and internet browser. I started preparing for meetings, lessons, and workshops with pen and paper instead.

This process slowed me down, quieting the ideas that normally pinballed through my head and dulling the temptation to reach for technology. Without AI-generated feedback, my first drafts were messier but far more authentic, genuinely capturing my ideas with much greater precision—a reminder of the value of process over product. And as a result, I’d managed to create a much firmer foundation for my own thinking. My pace and productivity undoubtedly slowed, but it felt like I was engaging in deeper work. Instead of asking ChatGPT for advice, I’d have to consult real human beings and wait for a response. This forced me to sit with my ideas longer, to wrestle with uncertainty and refine my own thinking. In many ways, the absence of instant answers became its own kind of teacher.

In a world increasingly dominated by speed and efficiency, the return to a more tactile workflow reminded me that some of the most valuable insights come not from rushing to get things done, but in the messy, unhurried act of giving something my full and undivided attention.

Toward Human-Centered Learning

No AI December wasn’t about rejecting AI (or anyone’s decision to use it), but about recalibrating its role in my life. Today, I’ve gone back to using these tools in ways that are far more intentional. My goal is to try to preserve the messy, creative core of what it means to be an educator: curiosity, critical thinking, and the grit and energy to solve problems without always resorting to shortcuts—technological or otherwise. My hope is that my students will do the same.

The challenge lies in designing classrooms where AI serves as a scaffold for innovation—enhancing, rather than replacing, intellectual rigor. 

Practice visible thinking: My experience reminded me of the widely applied pedagogy of “visible thinking.” Encourage students to lay out their thought processes, break down their reasoning, and explain how they arrived at their conclusions. For example, thinking routines like Think-Puzzle-Explore require students to articulate their questions, confront areas of confusion, and document their inquiry. This shifts focus from quick AI-generated answers to the value of productive intellectual struggle. By externalizing reasoning—through journals, annotations, or chat logs—students retain ownership of their learning, recognizing that true understanding often emerges through messy, human effort.

Celebrate student voice: Many students are using AI to improve their writing for them, rather than with them. This often results in AI stripping their work of all nuance, criticality, and voice. Let students know you want to hear what they have to say in their own words. Consider including “voice” in grading criteria and assigning low-stakes writing opportunities with ongoing formative feedback to help elicit more authentic expression. Engaging students with informed and opinion-driven writing can encourage them to articulate their own ideas and see the value in thinking for themselves.

Think before you prompt: We know that trying to solve a problem before being given the solution leads to better learning outcomes. Don’t allow brainstorming with AI to be the first thing your students do. When building foundational skills, consider having students occasionally unplug from technology entirely—drafting essays in class by hand or brainstorming on paper. 

Or ask students to journal or document their ideas prior to consulting with AI. Developing this habit is challenging but essential. You could even suggest an “AI delay” and ask students to spend 10–15 minutes (or more) generating ideas independently before turning to AI for additional support.

Unplug and reconnect: Challenge yourself and your students to embark on an AI detox. Step away from AI tools where possible, and rediscover the power of human creativity and independent thought. Start a journal to document the journey, exploring questions like “How does it feel to rely solely on my own intellect?” or “What challenges arise when AI isn’t there to assist?” This activity can foster resilience, raise self-awareness, and cultivate a deeper appreciation for the human capacity to think, create, problem-solve, and innovate.