Reimagining Middle School Education: The Synergy of AI and Montessori Principles

Next-Level AI Curriculum Development with Sam Bobo

Generated by DALL-E 3 via Chat-GPT 4 Turbo

Reading Time: 14 Minutes

What Will You Learn:

• Understand how AI can be seamlessly integrated into project-based learning experiences to create immersive, interdisciplinary educational opportunities that promote student autonomy and real-world engagement.

• Recognize the potential of AI to revolutionize middle school education by aligning with powerful pedagogical approaches like Montessori education, amplifying student choice, inquiry, and responsibility.

• Gain insights into how middle school can serve as a focal point for developing AI literacy in K-12 education, empowering students to become active participants in shaping the future of AI technologies.

---

Introduction

For many of us, middle school was one of the most challenging periods in our lives and educational journeys. During this time, our bodies are in active rebellion against themselves, undergoing rapid physical changes that can feel awkward and uncomfortable.

The social stakes of every success and mistake seem magnified through an ever-expanding sense of self-consciousness, as peer relationships and the need for acceptance take on paramount importance. Academically, the curriculum drills down into highly differentiated disciplines, often with little indication of interconnectedness or real-world relevance, leaving students struggling to find meaning and motivation in their studies.

However, over the past several decades, educational pioneers and innovators have entered the middle school arena, developing groundbreaking programs that offer students immersive, interdisciplinary, and project-based learning experiences. These innovative approaches have energized students, helping them overcome the institutional, societal, and interpersonal barriers that often hinder sustained academic growth and development during this critical period.

In my own career, I had the privilege of being trained as a Montessori Adolescent Guide and teaching on a suburban Erkinder campus. Erkinder is a philosophy developed by Maria Montessori late in her career, which regards adolescents as "social newborns" and emphasizes their need for meaningful work and independence as the context for instruction in traditional disciplines.

At our school, we put this philosophy into practice by engaging students in the cultivation of our own fruits and vegetables and the operation of a small business. These real-world, hands-on experiences served as the launching pad for our humanities, mathematics, and science units. By grounding academic learning in practical, purposeful activities, we aimed to help students see the relevance and interconnectedness of different subject areas, while also fostering their sense of autonomy, responsibility, and community.

As I reflect on this formative experience in my teaching career, I am struck by the ways in which the principles and practices of Montessori education align implicitly with the transformative potential of artificial intelligence in the classroom. When handled by skilled practitioners, both Montessori education and AI integration can amplify student choice, autonomy, inquiry, and real-world engagement.

In today's article, we will explore the unique potential of middle school as a pivotal point of sustained contact with AI for the current generation of students, positioning it as a potential accelerator for the adoption and integration of AI across K-12 education and the broader educational sector.

---

Guest Collaborator: Sam Bobo

I am thrilled to collaborate on this piece with Sam Bobo, a visionary leader in the field of Conversational AI and Generative AI technologies. As a Principal Product Manager at Microsoft and former IBM Watson specialist, Sam brings a wealth of expertise and passion to the realm of AI-powered innovation. Through his influential newsletter, Speaking Artificially, Sam shares independent insights and strategies for harnessing the transformative potential of Conversational AI.

As we delve into the possibilities and implications of AI in middle school education, I am confident that Sam's contributions will provide invaluable perspective and depth to our exploration. His commitment to empowering people to innovate, understand, and transform the world through groundbreaking AI technologies aligns perfectly with our shared vision for the future of education.

---

Middle School Curriculum: Inquiry, Autonomy, Real-World Engagement

The middle school curriculum continues to build upon the foundational pillars of a child’s broad, well-rounded, education that provides students with multiple, interconnected lenses in which to view the world around them. Students read a variety of literature to analyze the author’s diction and point of view, engage in algebraic and statistical mathematics, explore life sciences and the core tenants of scientific methodologies, delve into historical themes, and learn expression through the arts. 

Layered on top via the application of the Montessori Adolescent Guide interweaves these disciplines and places a spotlight on the student learner, following their unique individual needs and promoting both Socratic and collaborative dialog to solidify and build upon the complex subject matter. Furthermore, what delights me most about this methodology is the emphasis of project-based learning that establishes a deep connection with the material.

A Montessori approach imbued within a Middle School curriculum is prime for reinvention with AI as a primary tool. AI catalyzes the intellectual curiosity of student learners and enables a personalized lens on the learning material unscalable by teacher practitioners. Let’s dive into some of the characteristics of AI that unlock such a pedagogical approach and after, will provide an example of a Montessori Adolescent Guide project that could be implemented as a template for schools to mimic. 

The three GenAI characteristics we will explore include:

(1) Similarity search

(2) Medium transformation

(3) Conversational dialog

---

Similarity Search

Generative Artificial Intelligence (“GenAI”) employs a method called similarity search inside its vast corpus of training. When ingesting data, GenAI models tokenize, or segment, the textual input. These tokens can be individual words or logical portions of words that the model can understand. Post-tokenization, the data source gets vectorized. In contrast to humans who can comprehend only three dimensions — height (up/down), length (side-to-side), and width (forwards/backwards) — computers can comprehend n-many dimensions. 

Afterwards, the model undergoes a computationally-heavy training to compute parameters and finishes with a complete Foundation Model. What is most powerful about LLMs, in the context of the Montessori approach, is the vectorization, primarily through similarity search. Similarity is a basic vector computation in mathematics that computes the distance between two vectors. Should vectors be close together in n-dimensional space, they are similar. What does that mean for Montessori?

The Montessori Adolescent Guide approach to Middle School learning emphasizes the individual learner and their interests. Should a learner be highly interested in sports, for example, and grapple with complex topics such as photosynthesis — the process plants undergo to transform light into glucose sugar. GenAI could explain the Calvin Cycle in a soccer metaphor as:

This is the “Midfield Play” of photosynthesis. It happens in the stroma of the chloroplast. During the Calvin cycle, carbon dioxide (CO₂) gets fixed into glucose using ATP and NADPH. It’s like players passing the ball around, setting up the perfect shot.

Students are then able to expand learning of complex topics in familiar domains, boosting comprehension and retention. This application can expand to all student interests

---

Medium Transformation

Traditional Conversational AI capabilities, ones that emerged during the last era of AI circa 2014, and GenAI include pivotal services that perform medium transformation : Text-to-Speech, Speech-to-Text, Text-to-Image, language detection and translation. These transformations involve building models using recorded audio, and orthographic (transcribed) text for the speech related services and diffusion models for those producing visuals of any nature. 

Students can be categorized in four types of learners: auditory, visual, kinesthetic, and reading/writing. Through Artificial Intelligence as a tool, students are able to absorb context and produce content through the learning modality best suited for their comprehension. Should text need to be read out loud, employ text-to-speech, or should students need visuals to accompany text, utilize video or image generation. While these approaches are not perfect, they do expose a massive upside for students as the technology advances further.

---

Conversational Dialog

While there are certainly drawbacks to anthropomorphizing AI and not within the scope herein, a Conversational interface does provide a sense of comfort to those engaging with AI. For example, society largely prefers to engage with offerings such as ChatGPT, a conversational interface, than simply writing prompts in a tool of choice. Ever-expanding context windows being released with each iteration of publicly available models allow for deeper and deeper conversational turns as these systems are able to preserve ever-increasing memory of the conversation. 

These conversational systems open the door to Socratic dialog and provide a comfortable and familiar interface in which to conduct research, probe deeper on subject matter, or even ask for the subject-matter transformation. Most dialog in traditional systems are pre-constructed, with conversational designers postulating all possible use cases, routes, and phrasings for particular inquiries. GenAI interfaces, on the other hand, probe into the vast “knowledge” source of its corpus and use probability and statistics to generate content in a conversational manner and using sophisticated summary techniques to create a more dynamic interaction.

---

Case Study: “Flavors of the World” Exploration

In the introduction of this piece, Nick alludes to the cultivation of fruits and vegetables at the Erkinder school as a pedagogical tool for exploring the humanities, mathematics, and science. For our example, let’s expand that project into the culinary arts, whereby students will research and cook (with the assistance of a teacher or parent) up, and reflect upon an international cuisine. 

While I do not assume students have access to commercial kitchens at school, the example will display applicability into geography, language acquisition, religion, cultural anthropology, writing skills, science, and mathematics — all infusing Artificial Intelligence into all aspects of the project-based curriculum as a co-collaborator and tool.

Subject: “Flavors of the World” Exploration

Prompt: Choose a cuisine from anywhere in the world. Research its history, ingredients, and traditional cooking techniques.

Tasks:

• Select a specific dish from that cuisine.

• Virtually “Purchase” ingredients for the dish based on local habits

• Prepare the dish at home or in a cooking lab.

• Write a brief report explaining the cultural context and significance of the dish.

• Add your recipe and findings to the class cookbook.

Recipe Discovery 

Breaking apart the initial prompt, students would first brainstorm possible cultural dish ideas. Brainstorming could be in the form of self-introspection such as a traditional family recipe or generating a list of possible global cuisines. Students could use a chat-based interface to describe flavor profiles or components based on favorite foods as part of the dish, or even create a list of cultural dishes from a particular region and break down ingredients from there.

A student may start with a rudimentary understanding of global cuisine based on what their parents exposed them to in the home. Middle School students are likely vocal about favorite ingredients within dishes. The first interaction with an Artificial Intelligence tool may be a chat-based application powered by a Large Language Model (think ChatGPT or Copilot) by prompting it with: 

Please share 5 cuisines from different global cultures that use some of my favorite ingredients including {ingredients}. Please omit results that include {restrictions}.

For example: I am a vegetarian exploring global cuisines. Please share 5 cuisines from different global cultures that use some of my favorite ingredients including garlic, spinach, and cheese

What resulted included:

• Spanakopita: Spanakopita, a delightful Greek spinach pie, combines perfectly flaky layers of phyllo dough with a comforting filling of spinach and feta cheese. Let’s explore the key ingredients and how to make this savory delight: It’s both nutritious and flavorful! 

• Indian Cuisine: These flavorful bowls feature curried chickpeas, brown rice, and garlicky spinach. Top them with fresh cilantro and green onion for added freshness. It’s a wholesome and satisfying meal that’s perfect for meal prep too! 

• Spinach and Cheese Enchiladas: These enchiladas are filled with sautéed spinach, garlic, and your favorite cheese (vegan or dairy-based). Top them with a zesty tomato sauce and bake until bubbly and delicious!

• Vegan Cannelloni with Spinach Ricotta Cheese: This Italian-style dish features pasta shells stuffed with a flavorful vegan spinach cream ricotta cheese. It’s baked in the oven with a creamy tomato sauce and dairy-free cheese on top. Garnished with fresh basil, it’s a delightful twist on the classic cannelloni

The next turn in the conversation may be to solicit the ingredient list and recipe, say for Spanakopita, generating the ingredient list below and a step-by-step guide to preparing Spanakopita (not included for brevity).

Created by Midjourney

• Spinach: You’ll need 16 ounces of frozen chopped spinach (thawed and well-drained).

• Fresh Parsley: 2 bunches, finely chopped.

• Onion: 1 large yellow onion, finely chopped.

• Garlic: 2 cloves, minced.

• Quality Feta Cheese: Approximately 10 ½ ounces, crumbled.

• Dried Dill Weed: 2 teaspoons.

• Eggs: 4 large eggs to bind everything together.

• Olive Oil: 2 tablespoons for sautéing and flavor.

• 16-ounce package of Fillo Factory Organic Dough (contains 4 pastry sheets).

• Olive Oil: About 1 cup (or as needed) to brush between phyllo layers.

Virtual Shopping & Historical Significance

Next, the student would purchase the ingredients virtually from the local grocery store. For illustrative purposes, assume the learner belongs to a school with a high technology budget to equip students in a computer lab with virtual reality headsets. Virtual Reality, while an emerging technology, provides an immersive experience for the user. Game companies such as Ubisoft are highly skilled in recreating historically accurate representations of cities, creating realistic NPCs (non-playable characters) through Generative AI, and highly convincing acoustics. As an aside, Ubisoft’s historical digitalization of Notre Dame helped architects rebuild the historic church after the devastating fire in 2019.

Created by Midjourney

Learners would enter the immersive world, say in Athens, and walk to the local grocery store. Within the virtual environment, students would listen to ambient chatter in Greek, traverse the historic streets of Athens, and view a simulation of local life in the area. They could approach NPCs and engage in conversation to ask questions, using machine translation to convert the students native language, say English, into Greek and vise-versa to expose them to the language. 

This back-and-forth dialog could delve more into the Greek culture and the significance behind the dish they were creating. After speaking to a local, the student may discover that Spanakopita is representative of festive gatherings that draw family, friends, and community together, a pillar of the Greek culture. One such holiday is Names Day (Ονομαστική Εορτή) where Greek children hold a celebration on the day associated with the Saint they were named after, a common tradition in Greece stemming from the Orthodox Church.

Students who are unfamiliar with Name Day may ask for the tradition to be described in a different, more familiar topic, such as music. They may learn that the calendar itself is like a musical score with a day corresponding to a note. People are represented as instruments in the symphony and play a unique note with their instrument (their name). On Names Day, the musician gets a solo, imbuing the emotion of the Saint’s significance into the celebration of the name (instrument and notes).

Write Up

After immersing themselves in a virtual world, gaining the historical significance of the dish through conversation and experimentation, virtually purchasing the food with local currency, and having casual conversation, its time to construct the final deliverable: the recipe book. Students could use text-to-image (e.g DALL-E) to create a picture of the dish using descriptive terminology, copy the recipe received from the chat interface, and display a level of mastery through a self-guided reflection written up alongside the recipe. 

---

Sam’s Conclusion

The “Flavors of the World” Exploration project tied together language acquisition, geography, mathematics (currency conversion), writing proficiency, science (unit conversions within recipes), and art (image generation) seamlessly together in an AI-powered experience (along with technological schools as a given — using AI). While this level of immersion and coordination with Artificial Intelligence is simply aspirational at the present time, it lays out the foundation and roadmap for schools and educators to apply AI in a safe and purposeful way using the Montessori Adolescent Guide as a pilot for the exploration. Maria not only believed that project-based learning was an essential aspect of pedagogy, but also sought real world application of such projects, particularly ones that focused on agriculture to teach children about the foundational elements of the world.

___________________________________________________________________________

Nick’s Takeaways

As we conclude this exploration of AI's transformative potential in middle school education, I want to express my deep gratitude to Sam Bobo for his extraordinary contribution to the ongoing dialogue at Educating AI. Sam's expertise, passion, and visionary leadership in the field of Conversational AI and Generative AI technologies have been instrumental in shaping our understanding of how these powerful tools can be harnessed to revolutionize education.

Our collaboration on this piece represents a next-level approach to AI pedagogy, one that goes beyond simple prompting guides, lists of pros and cons, and surface-level explorations of potential use cases and tools. Instead, we have delved deep into the principles and practices that underpin truly transformative educational experiences, drawing on the wisdom of time-honored pedagogical traditions like Montessori education to illuminate the path forward.

By aligning AI with these powerful pedagogical approaches, skillful practitioners can amplify student choice, autonomy, responsibility, inquiry, and real-world engagement in ways that were previously unimaginable. The "Flavors of the World" Exploration project we have outlined here is just one example of how AI can be seamlessly integrated into a project-based learning experience that ties together language acquisition, geography, mathematics, writing proficiency, science, and art in a way that is both immersive and deeply meaningful for students.

As we have seen, the combination of AI and these pedagogical traditions has the power to ignite a revolution in middle school education. By positioning middle school as the locus for AI literacy in K-12, we can create a space where students are not just consumers of AI technologies, but active participants in their development and deployment. Through sustained engagement with AI in the context of real-world projects and inquiries, middle school students can develop the skills, knowledge, and dispositions necessary to thrive in an increasingly AI-driven world.

It is time for educators to move from contemplation to action. As Ethan Mollick challenges in his new book Co-Intelligence, we will only discover amazing new uses of AI by actually using AI. We invite other educators to develop projects like our own and share them in the comments. We need to start building up a shared database of resources rather than keeping our experiments to ourselves. Only in this way will we fashion a safe, engaging, and innovative educational space for this and the next generation of students.

In closing, I want to reiterate my deep appreciation for Sam Bobo's contributions to this piece and to the broader conversation around AI in education. His insights and expertise have been invaluable in helping us chart a course towards a future in which AI is not just a tool, but a catalyst for transformative learning experiences that empower students to become the creators, innovators, and leaders of tomorrow.

As we continue to explore the possibilities and implications of AI in education, I am confident that our collaboration will serve as a model for the kind of deep, thoughtful, and impactful work that is necessary to realize the full potential of these technologies.

Nick Potkalitsky, Ph.D.