How to Thoughtfully Use AI to Create Meaningful Lessons

From fears about academic dishonesty to declining foundational skills and increased screen time, there’s no shortage of concern about how students use artificial intelligence (AI). While I share many of those concerns, I’ve discovered how AI can also be a force for good if we use it with intention. In my experience as a math and computer science teacher, the most effective use of AI tools hasn’t come through student-facing platforms or shortcuts. Instead, AI has become a behind-the-scenes assistant that helps me save time, spark ideas, and design more engaging, collaborative learning experiences that preserve the human element in my classroom.

Computer science is a rapidly evolving subject closely tied to shifting technologies. I want my students to learn current, relevant skills. Staying ahead of the curve often means building new units on unfamiliar topics. In the past, searching for ready-made resources online or jerry-rigging my own cobbled-together activities felt disjointed and unsatisfying. My efforts often had an amateurish feel that students could sense, and I wasn’t proud of that.

Create Scaffolded and Refined Lesson Plans

AI chatbots changed my patchwork results. Rather than relying on apps made specifically for teachers, I’ve opted to use more unfettered commercial AI tools like ChatGPT, Gemini, and Claude. I’ve found that the paid versions are well worth it. Where AI was once surprisingly clumsy and inept at even basic algebra, the new models optimized for coding and logic are astonishingly good.

For example, while developing a unit on computer networking (a topic I knew little about), I fed ChatGPT my broad goals, specific classroom constraints, and a few initial ideas. I started with a general prompt:

• “Create a unity plan for a high school introductory computer science class on networking. Include as many interactive activities as possible.”

From the results of that prompt, I added constraints and features to tailor the results to my needs. I specified that the unit needed to be project-based, couldn’t rely on the availability of specialized networking hardware, needed to run on just Chromebooks, and should build on students’ basic Python coding skills:

• “Suggest activities that involve some basic Python programming.”
• “How do I set up a lab sandbox environment for students in a high school computer science class to learn about the fundamentals of networking using Chromebooks?”
• “Include hands-on exercises on a Linux server where possible.”

The end result was a tailored unit plan with scaffolded lessons and activity suggestions, including ready-to-use materials like Google Colab notebooks. What once felt like an impossible task became one of the highlights of the year.

This approach has proven to be most constructive when I begin with a rough set of notes about what I want the unit to accomplish—allowing the AI to suggest the fine-grained structure. Standards alignment can come later; I’ve learned that early drafts are more useful when they’re written with a focus on what students will experience front and center. While iteration and refinement are still required, the time saved and the overall improvement in depth and coherence make it worthwhile.

Save Time and Support Collaborative Learning

Even at the lesson level, AI has helped me shift from teacher-centered models to more collaborative learning. The obstacle isn’t a lack of ideas; it’s time. Whether designing a vertical whiteboard session using Peter Liljedahl’s Building Thinking Classrooms in Mathematics strategies or preparing a review activity for the next day’s introductory computer science (CS) class, the time needed to generate the critical mass of problems necessary to support collaborative activities with the right level of focus and variety is significant.

For instance, creating a tarsia puzzle or scavenger hunt might require 30 different but related problems. It’s not hard work, but it’s tedious and time-consuming, especially when balanced against grading, planning, and other responsibilities. Through prompt engineering, I’ve trained AI to generate these problem sets quickly and in multiple formats, including graphs and code snippets. This has enabled me to implement collaborative activities—Quizlet Live games, tarsia puzzles, and scavenger hunts—with much greater frequency and impact.

Beyond generating lesson content, I’ve used AI to help me create student-facing materials. During the Advanced Placement test review in my calculus class, we needed to drill multiple-choice problems. Rather than revert to worksheets, I asked ChatGPT for ways to make the experience more interactive. One suggestion was an escape-room-style Google Form with self-checking elements. While AI can’t directly create a Google Form, I’ve discovered that it can write a Google Apps Script (basically JavaScript) that builds the form automatically. That saved me hours. I used the extra time to improve the activity’s aesthetic appeal with images and videos, transforming a potentially monotonous review into something that my students and I found to be energizing and fun.

Formatting support is another unexpected benefit. Since AI chatbots typically return results in markdown format, I can paste content directly into Google Docs and still preserve its structure. I’ve used this to generate handouts, Python exercises, and even LaTeX-formatted materials for complex math concepts. I’ve also asked ChatGPT to create Google Sheets rubrics that are formatted for direct upload to Google Classroom. On the CS side, I’ve used AI to produce Python-based Jupyter notebooks, which can be imported directly into Google Colab to create interactive and dynamic student coding environments.

Fill in Gaps, Create Cohesive Lessons, and Boost Engagement

In my computer science classes, where screen use is inevitable, AI tools have allowed me to design tasks that are far more relevant, collaborative, and project-based, tailored to our specific classroom context. My entire CS curriculum is now more coherent and consistent, with AI helping to bridge gaps in my background and experience. In math, I can implement various collaborative activities more regularly, and my Building Thinking Classrooms whiteboard sessions have become more sustainable and robust.

I also integrate online activities more intentionally, always intending to deepen collaboration and boost student engagement. Paradoxically, since I’ve incorporated AI more deeply into my workflow, my students have been able to spend less time passively using screens and more time actively engaging with each other and with the content. By freeing up my time and expanding what I can realistically prepare, I’ve created more student-centered, socially interactive classroom experiences and activities that would have been out of reach without AI’s support.

This is a pivotal moment for educators and society at large. As we figure out how to teach students to use AI responsibly as a tool and not a crutch, we also need to consider how we, as teachers, can harness it ourselves first. Used thoughtfully, AI can enhance and not replace our practice. It gives us time to be more creative, responsive, and connected with our students. In my classroom, AI hasn’t made things less human. It has made them more so.