Remembering SAMR: Moving AI Integration from Substitution to Transformation
Last week, I had a conversation with a longtime friend and computer science teacher who reminded me: "Don't forget SAMR." We had been discussing teacher uses of AI and how many seem satisfied if it merely saves them time. "If all we're doing with AI is substitution and augmentation," he told me, "then we're falling well short of what AI is capable of."
I must admit, I hadn't really thought of the SAMR model for a while. About a decade ago, I frequently ran "SAMR and iPads" workshops and had met SAMR's developer, Dr. Ruben Puentedura, in 2013. Suddenly, I was reminded of how the SAMR model provides a valuable framework for reflection, particularly in the context of AI integration in education.
The SAMR model outlines four levels of technology integration: Substitution, Augmentation, Modification, and Redefinition. At the Substitution level, technology acts as a direct substitute, with no functional change, like digital versions of paper worksheets. Augmentation also involves substitution, but with functional improvement. For instance, an English teacher might have her students use Grammarly to get automated suggestions on their grammar and spelling as they write. While the AI helps catch errors and provide suggestions, the core task of writing and the teacher's role in providing feedback remain largely the same.
(Wikimedia image) Though often presented as a ladder, the SAMR model is more like a spectrum because technology integration activities can move back and forth through the spectrum.
Modification is the starting point for AI's transformative potential. It involves a significant task redesign as teachers bring about substantial changes to the learning process. For example, students might use Desmos to create visual models of complex real-world situations, like population growth or climate change, and then use ChatGPT to generate insights, predictions, and potential solutions based on the models. This significantly modifies the traditional approach to mathematical modeling, making it more interactive, iterative, and relevant to real-world problems.
Redefinition involves the creation of new tasks, previously inconceivable without the technology. Imagine, for instance, a science class where students team up with AI agents that crunch massive datasets, run simulations, and generate novel analysis in real-time, while also explaining their reasoning and responding to student queries. This level of human-tech collaboration in the classroom would be unfeasible without sophisticated AI systems.
The recent Walton Foundation report findings, which includes how teachers are using AI at the Substitution and Augmentation levels, also got me thinking about SAMR. The report indicates that AI is primarily being used at the lower levels of the SAMR model. Teachers are mainly using AI to generate ideas for classes, prepare instructional materials, create student worksheets and quizzes, and automate grading. While these uses can save teachers valuable time, they don't fundamentally change the nature of the students' learning experience.
The most popular responses to the question "Which of the following are ways you have used AI chatbots for your job?" particularly struck me. None of the responses involved putting technology into the hands of students. It's virtually impossible to move AI integration beyond substitution and augmentation if students are not actually using it.
I am excited to announce that I will be releasing my new book AI Tools & Uses: A Practical Guide for Teachers at the end of the month! The book is full of top tools, implementation strategies, spotlights, and case studies to help teachers integrate artificial intellignce into their instructional practices.
All subscribers to this newsletter will receive a free sample of the book, including three chapters! Encourage your colleagues to sign up for this newsletter before June 30, so they too can receive a free sample!
Modification and Redefinition
So, what might AI integration look like at the Modification and Redefinition levels? Here are a few ideas:
Mathematical Modeling with AI (Modification): Incorporate Desmos and ChatGPT into lessons on mathematical modeling. Students can use Desmos to create visual models of complex real-world situations, like population growth or climate change. They can then use ChatGPT to generate insights, predictions, and potential solutions based on the models. The AI can help students analyze the relationships between variables, make data-driven decisions, and communicate their findings effectively. Rather than simply using Desmos as a substitute for hand-drawn graphs or calculations, the integration of ChatGPT fundamentally changes the nature of the activity. Students can now engage in a more dynamic, exploratory process, using the AI to generate insights and refine their models in real-time.
Virtual Language Immersion with AI (Redefinition): AI-powered platforms like Talkpal can provide foreign language students with immersive, interactive speaking practice in realistic scenarios. Students can engage in open-ended conversations with AI language tutors, receiving instant feedback on their pronunciation, grammar, and fluency. The AI adapts to each student's proficiency level, offering personalized challenges and support. This redefines language learning by enabling authentic, contextualized practice that would be difficult to replicate without AI technology. It creates a transformative language learning experience that would be virtually impossible without AI.
Socratic Seminars with AI (Redefinition): In this scenario, students engage with an AI-powered Socrates about a complex text like Plato's Republic. The AI takes on the role of Socrates, posing thought-provoking questions, challenging assumptions, and providing contextual information. Students engage in critical thinking, construct arguments, and explore philosophical ideas in a simulation with Socrates, which would be impossible without the AI. This redefines the concept of a Socratic seminar, making it more immersive, interactive, and accessible. The AI becomes an active participant in the discussion and guides student exploration of complex philosophical concepts, enabling an entirely new kind of interaction that would be inconceivable without the technology.
Adaptive, AI-powered science simulations (Modification): In a traditional science class, students might perform a set of predefined experiments or simulations to explore a particular concept. To integrate AI at the Modification level, the teacher could use an adaptive simulation platform that adjusts the parameters and complexity of the simulation based on each student's performance and understanding. As a student manipulates variables in a physics simulation, the AI could analyze their inputs and results, providing real-time feedback and adjusting the difficulty level. If the student is struggling, the AI might simplify the scenario or provide additional guidance. If the student is crushing it, the AI could introduce more complex challenges or prompt the student to explore related concepts. The AI fundamentally transforms the nature of the science experiment or simulation, creating an adaptive learning journey tailored to each student's needs.
As you can imagine, moving from substitution to transformation is not without challenges. It requires a significant shift in mindset, willingness to experiment and take risks, and an understanding of both the potential and limitations of AI. The AI-integration activity itself might also raise important questions about data privacy and algorithmic bias.
Teachers need support to effectively use AI and engage students in conversations about responsible AI use. We must remember that AI is not a solution in and of itself; it is a tool whose impact depends on how it is used. AI should augment and enhance human capabilities, not be a substitute for them. The most transformative uses of AI will be those that empower teachers to create more engaging, personalized, and effective learning experiences for their students.
As we move forward, let’s keep the SAMR model in mind. Let’s strive not just for substitution, but for transformation. Let’s use AI not simply to automate certain tasks we do, but to reimagine what's possible. And let us always keep our students at the center, letting them use AI in ways that support their learning, growth, and success.
AI Tools and News
New AI tools for Google Workspace for Education - Google announces (again) Learning Coach, a custom version of Gemini powered by LearnLM, as well as other developments.
TeacherServer - an AI teaching assistant
Candide - “The AI school open to everyone.”
Best Sites for Online Tutoring and Teaching - Tech&Learning
5 Reasons The OpenAI and Apple Partnership Intrigues Me As An Educator - Tech&Learning
AI Plagiarism Considerations Part 1: AI Plagiarism Detectors -
Illuminate - experimental tool to foster learning through published academic papers
Claude 3.5 Sonnet is here - Anthropic
Beyond Chatbots: How Educators are Using Claude's Artifact Tool in Sonnet 3.5 - Stefan Bauschard
Creating AI Tutors that Don’t Hallucinate - Tech&Learning
Dream Machine - transform text and images into high-quality, realistic videos
Runway Launches Gen-3 AI Video Model - Runway
Gmail’s Gemini AI sidebar and email summaries are rolling out now - The Verge
Open AI Launches ChatGPT for Mac - OpenAI
Amazon reportedly working on new AI chatbot to compete with ChatGPT - SiliconAngle
DeepMind’s new AI generates soundtracks and dialogue for videos - TechCrunch
Toys “R” Us riles critics with “first-ever” AI-generated commercial using Sora - ars technica