Visualizing History: Reading Comprehension Through AI-Generated Images

A guest post from Ned Courtemanche

Today’s post comes from Ned Courtemanche — History Department Chair at McDonogh School and Curriculum Design Specialist with AI Literacy Partners. Over the past two decades, Ned has driven educational innovation as a teacher, administrator, and educational consultant. In his current role outside of Baltimore, Ned approaches Edtech with equal parts skepticism and curiosity, having actively experimented with AI in his classrooms for the past two years.

He has served as a content creator for Flint AI, published a blog series exploring teacher-driven AI as a force multiplier for critical thinking, and facilitated AI workshops at events like the Association of Independent Maryland & DC Schools (AIMS) Technology Summit. Ned is deeply committed to expanding educator access to AI tools and ensuring that teachers determine how this disruptive technology reshapes education.

Today, he shares his breakthrough approach to addressing one of education's most persistent challenges: declining reading comprehension scores. His "Visualizing History" curriculum transforms decades of research on visualization techniques into a practical, scalable classroom solution using AI image generation. The complete lesson plan is downloadable for free via our Assessment Redesign Services page, and this post tells the story of its real-world implementation.

You'll find Ned’s plan alongside our ELA packet on The Great Gatsby, Wess Trabelsi's chemistry mystery powders lesson, and our soon-to-launch math activities. Just head to the Assessment Redesign Services page and click Download under the History section.

In the meantime - enjoy your experimenting.

And let’s grow together,

Mike

---

Visualizing History: Reading Comprehension Through AI-Generated Images

By Ned Courtemanche

We've all heard that reading comprehension scores are in a tailspin. From PISA results to national assessments, the data tells a consistent story: students struggle to extract meaning from text more than ever before. Meanwhile, decades of research have highlighted a technique with the power to pull back on the stick and enhance reading comprehension: visualization. Allan Paivio's Dual Coding Theory, John Hattie's meta-analysis, and countless classroom observations all point to the same conclusion: when students create mental images while reading, comprehension improves dramatically. The challenge has always been implementation. How do you scale visualization exercises consistently with five sections of over a hundred students and limited prep time? It's a logistical and time-intensive nightmare worthy of Hannibal crossing the Alps that most teachers simply avoid.

Enter AI image generation. As I explored in a previous blog post, over two years of experimentation in the classroom has led me to believe that AI can be a force multiplier for critical thinking with the ability to enhance core skills like reading comprehension. AI image generation suddenly makes visualization techniques not just practical but powerful. Let me show you what this looked like just recently for my sophomore Modern World History students with a lesson that any teacher can implement tomorrow.

The Lesson That Works Everywhere

The "Visualizing History" curriculum that I'm sharing with AI Literacy Partners is deceptively simple: students read a text carefully, identify four key passages, then use AI to generate images that reflect their understanding. We ask students to think like movie directors, crafting specific prompts, evaluating the results, refining their approach, and ultimately creating a visual storyboard of their reading. The entire process takes one class period and transforms passive reading into active interpretation.

This approach offers tremendous value for teachers. It works with ANY reading: primary sources, scientific articles, literature, even math word problems. The curriculum adapts to any grade level and requires no AI expertise. In fact, teachers new to AI will find this an ideal entry point. All materials are provided free through AI Literacy Partners, and learning alongside your students becomes a powerful teaching moment in itself. The multi-level impact extends beyond reading comprehension to include content acquisition, AI literacy, and essential soft skills like patience and problem solving.

For many students, this activity also reveals the gaps in their AI fluency. My current sophomore baseline data was eye-opening: 74% rarely used AI (once a week or less) for personal reasons, and exactly zero mentioned using it for visual creation or creative expression. Their top uses? Study questions, summarization, and academic clarification. These students who've grown up with technology aren't as AI-literate as we sometimes assume. All the more reason to run this engaging lesson that addresses both teacher needs for scalable visualization exercises and student needs for what is fast becoming an AI-driven world.

Phase 1 & 2: Visualizing Text Through Directed Practice

Picture this: Modern World History, second period, with our study of decolonization movements leading us to India and Gandhi's Salt March. I introduce the day's objective - using AI to visualize historical texts - and hear only one or two audible groans. In the lexicon of sophomore reactions, that's practically a standing ovation!

I pull up our AI image generator and type: "A medieval castle." The result is predictably vague: a generic fortress that could exist anywhere, anytime. "Now how would a Hollywood movie director approach this?" I ask, guiding students to consider Characters, Scene, and Style in their prompts before evaluating the output and revising, what we call the C.S.S.R. framework.

The class builds the prompt together: "Who's there? A knight returning from battle. When? Sunset. Style? Make it look like a Renaissance painting." We enter the revised prompt, and the transformation is immediate. The generic castle becomes a specific moment, a story, an interpretation. "Remember," I emphasize, "we're not aiming for perfect historical accuracy. We're after enhanced understanding and informed artistic interpretation."

After running this activity dozens of times, I've learned that visual creation consistently engages students more than some of our other AI-empowered activities. Students are often initially surprised and excited to find themselves with the power to create compelling visuals, and a few always rush ahead into the next phase of the assignment. But AI image generation can be finicky, almost like GPT 3.0 - it's one of the few AI spaces where human input genuinely makes or breaks the output. No hand-holding from an overly helpful chatbot here. Students will need to really dig into the reading, hone their prompts with the C.S.S.R. framework, and draw from a range of soft skills to effectively navigate what's to come.

Phase 3: Where Reading Comprehension Comes Alive

The assignment seems straightforward: create four AI-generated images based on passages from their reading. What unfolds is beautifully messy and will require your "guide on the side" assistance, moving between students for those precious one-on-one interactions that makes teaching worthwhile.

As predicted, the first burst of my sophomore enthusiasm turns to frustration for some: "The AI isn't listening to what I'm telling it!" But by their fourth image, they’ve shifted perspective: "Chat had the same vision from the start, it just needed some tweaking." This progression from blaming the tool to recognizing their role in communication is one important takeaway, with roughly half of this class noting this breakthrough in later reflections.

AI image generation speed is also a common frustration. "It took a long time for the AI to actually generate an image and keeps making me answer follow up questions," one student noted. This is precisely why I distribute the AI Results Worksheet, it transforms those 1-2 minute wait times into productive reflection. Students record their prompts, note their thinking, and evaluate results while images generate. It's essential to both the immediate learning and their later metacognitive reflection.

Most students eventually recognize the transformative power of 'Style' prompts, which provide the AI with artistic parameters that unlock entirely new worlds of visual expression. You can always share this insight in Phase 1, but I prefer to wait and watch which students discover they can invoke "Van Gogh" or "dystopian" or "watercolor." One particularly creative sophomore this go-round asked, "What would this look like in a Bollywood movie?" That's when they realize they're experimenting with artistic interpretation of historical texts, not just following instructions.

Then there's what one sophomore called the "floating head brain flop", as the AI inexplicably generates Gandhi's head floating above the scene à la Salvador Dalí. AI and AI image generation are unpredictable by nature, but these are prime teachable moments for our students to develop deeper instincts for the challenges and opportunities of these tools. This particular student's later reflection - after revising to attach Gandhi’s head - captures the nuanced thinking that emerges in most every class: "Not exactly how I pictured the scene... Besides that, the image is close to perfect. AI is still bad for the environment and I don't think we should use it frequently, but this activity was cool."

Phase 4 & 5: From Frustration to Understanding

The share-outs and homework reflections revealed some interesting learning patterns from this group of sophomores. Fifty percent identified "being specific" as the key to working with AI. Another 30% recognized the necessity of "trial and error." One enthusiastic student recommended "using as many words as you can!" Not quite right, but I love the hustle.

The journey from frustration to competence (and sometimes back again) is an important part of the process. One student who complained during the activity that it was "rough, the AI isn't listening" later reflected: "I liked how chat had the same vision from the start, it needed some tweaking but overall it did a very good job." Another noted, "It's getting easier to work with AI, sometimes it's not always exactly what I picture." Of course, not everyone thrives. A handful tap out, frustrated by delays or wanting more control. As one student honestly wrote: "This was frustrating, AI doesn't listen with slow loading and processing." Fair enough, at least they understand the limitations. This meta-cognitive shift - grasping that quality AI output requires effort, iteration, and patience - is exactly what we're after.

But the deeper pedagogical win comes through clearly in reflections like this: "Personally, I enjoyed the AI images because it put concepts into fruition for me." Students who struggled technically still emerged with enhanced comprehension of the source material. The act of selecting passages, translating them to visual prompts, and evaluating results forces close reading in ways traditional comprehension questions rarely achieve. Perhaps most importantly, it has also sparked some of the best classroom discussions I've had in recent memory about the strategies and human experiences of decolonization movements. When students are inspired to immerse themselves in reading, to explore, question, and create, reading comprehension often takes care of itself.

Your Classroom, Our Turn

This approach exemplifies how AI can enhance proven pedagogical practices. For those of you still uncertain about AI integration, know that unpredictability comes with the territory - from impatient students to floating heads. The key is embracing this vulnerability and learning alongside your students while maintaining your role as the domain expert.

As Dr. Nick Potkalitsky argues in a recent blog post, the ‘cost of waiting’ for teachers to develop AI mastery before classroom use is far too high. The students aren't waiting, they need us to help them navigate these tools now, imperfectly but purposefully. The free curriculum available through AI Literacy Partners includes everything: lesson plan, student handouts, worksheets, even the Gandhi source text. You could literally teach this tomorrow.

This isn't disruption, it's Paivio with pixels, Hattie with hardware. We're taking research-validated techniques and making them scalable through technology. Yes, it's messy. That's where learning happens.

I'm thrilled to be partnering with AI Literacy Partners to bring more classroom-tested, teacher-proven approaches to educators everywhere. While Silicon Valley debates AGI, we're discovering what actually works in real classrooms. The future of education shouldn’t be shaped by big tech or politicians, but by teachers willing to experiment, fail, iterate, and share what’s pragmatically possible.