Back to School with AI, Part 1: Make Sense of AI
Three guiding questions, three principles, and seven things to try.
Welcome to Learning on Purpose! I’m so glad you’re here.
I have a lot to learn and figure out as I get this going. I’m not sure yet how often I’ll post nor which topics I’ll focus on, but I’m working towards finding a cadence so that you’ll know when to expect to hear from me. Thanks for bearing with me in these early days.
One thing I DO know: I wanted to start by writing about generative artificial intelligence (AI). I have a few posts planned for the coming weeks, but don’t worry, this won’t be an AI-only newsletter.
For me, the most important first step schools can take with AI is to help educators make sense of it.
In the workshops I’ve led on generative AI and assessment, and in the many conversations with schools I’ve had over the months since ChatGPT made its debut last year, I’ve found that educators, particularly teachers and staff who work directly with students, are experiencing a range of emotions about AI: skepticism, excitement, anxiety, resolve, and exhaustion. The sudden explosion of AI is just the most recent in a seemingly relentless string of challenges to the teaching profession since the onset of the coronavirus pandemic in 2020.
I recently left a great job at a great organization where I spent a decade working with teachers and administrators on using technology and learner-centered design to improve the student experience. For me, there are two essential elements to supporting educators in adapting to innovations in education: purpose and practice. We need to be clear about why an innovation is relevant to their work and to their students, and then we need to introduce some basic strategies that give them enough confidence to explore on their own. The goal is not to get educators to fall in love with an innovation, it’s to help them make sense of it so they can make informed decisions on their own.
In that spirit, below are questions, guidelines, and examples of ways I’ve used AI myself and with educators in order to make sense of it. This is by no means a comprehensive list: I would love for you to add your own ideas and suggestions in the comments.
Three Guiding Questions
After months of thinking and talking about and trying AI, I keep coming back to three questions:
What is our job as educators in a world being disrupted by AI?
How might AI help students think more deeply?
How might AI help educators become better designers and facilitators of learning experiences?
How we talk about AI matters. For example, I was moved by this English teacher’s reframing of AI from “cheating” to “assistance.” I prefer these questions to the more general and more abstract (and very common) “What should we do about AI?” How we frame AI for ourselves as educators will affect how we frame it for our students, and I believe strongly that how we talk to students about AI will have a lasting impact on how they see its role in their own lives.
Three Principles for Using AI
Start with what you know. I was a middle and high school English teacher for 12 years, so most of my practice with AI has focused on things like grammar and essay writing and literary analysis. Why? Because I was in a position to assess the quality of AI’s responses. My content knowledge has helped me craft prompts and follow-up questions to coach AI to produce better work.
Focus on process, not product. We should be worried about artifacts of work that AI creates for us and for students, but we should be equally aware of how it can support our learning and work processes. Consider three areas: efficiency (can I get things done faster with AI?), effectiveness (can it make my work better?), and innovation (can it help me process and act on new ideas?).
Prioritize dialogue over prompting. Not unlike those of human beings, AI’s first attempts are rarely that great. Fortunately, it responds well to coaching and feedback. I am encouraging educators to spend less time trying to craft perfect initial prompts and instead engage AI in more sustained dialogue: ask it to do something, then nudge it and redirect it and correct it several times. Pay attention to how it learns what you want and then quickly improves its work in surprising ways.
Seven Things to Try
For me, the best way to make sense of AI is to use it. These tools are powerful, but they are imperfect. They have enormous potentials and pitfalls for our world, but the tools available to most of us are accessible and easy to use. They are already doing amazingly sophisticated work in fields like medicine and technology and business, but they also have simple and practical applications in our day-to-day lives.
Open up a chatbot like ChatGPT or Bing or Bard or Claude and give a few of the below ideas a try. All of these tools are free to use (sometimes with limits). Do this with colleagues! This kind of work is wonderful to do in groups, first because it’s more fun and second because you can compare notes and share ideas in real time. It’s a way to normalize open, collaborative exploration of AI.
Create a Rubric. I have found AI chatbots to be particularly good at generating learning outcomes like competencies, standards, and performance levels for rubrics. Here’s a prompt I used: “Generate a rubric in table format for the competency ‘Communicate effectively with AI.’ Include five sub-competencies. Articulate four performance levels for each subcompetency: novice, improving, mastering, and advanced.” What would you ask AI to do after you saw this first attempt?
(Re)design Feedback. Input into AI an example of written feedback you have given to a student or a colleague (with any personal or identifying information removed). Ask it to give you feedback on your feedback. Ask it to reformat the feedback from narrative to bullet points (or vice versa) in order to highlight key points. Ask it to rewrite the feedback (be warmer, be more direct, be more specific, etc.). Ask it to look for bias. Ask it to generate follow-up questions the person receiving the feedback might have. Is it making your feedback better? Is the dialogue helping you think differently about your feedback?
Make Your Assignment an AI Experiment. Ask AI to complete some of the assignments you give students. When I was teaching, one of my favorite assignments was the Junior Profile: all 11th graders wrote a New Yorker-style profile of a person at work. They would spend six weeks identifying a subject, interviewing and observing them, then writing and revising. Here’s what ChatGPT generated for me in a few minutes. How did it do? What prompts would you give it to refine this initial response? What does this mean for the future of this assignment and how we collaborate with students on it?
Generate Questions. AI can be a question-generating machine. Ask it to create quiz questions for your students. Ask it to sequence those questions by difficulty. Ask it to pretend to be a teacher and quiz you on a subject. Go beyond student-facing questions: Have it generate questions for parent conferences or interviews or teacher evaluation meetings. I once asked ChatGPT what questions a new teacher might have for a principal during onboarding. How many schools have great answers to all of these questions? How might AI help us better anticipate what people need from us?
Create Models and Examples. Input one of your rubrics into a chatbot and ask it to create a sample of student work that aligns to the highest performance levels. Ask it to create an example that’s more middle-of-the-road. Ask it to create examples in a variety of voices or tones or topics. What does it take to get AI to produce models that you like? How might these AI-generated models be useful to your and your students for learning?
Practice with new ideas. This, to me, is one of the most exciting uses of AI: practice applying new ideas to our work. I asked ChatGPT to create a two-week unit on rhetorical devices that applied concepts of spaced and retrieval practice. After its first attempt, I asked it to make the summative assessment more performance-based. Here’s the whole interaction, which took about 20 minutes. Later on, I posed as a school leader and asked ChatGPT to design an interactive faculty meeting agenda introducing educators to AI that included a station rotation. In both cases I was struck by the psychological safety I felt: there was so little vulnerability in trying new things with AI as opposed to in front of students or colleagues. I appreciate that AI can be a place for educators (and students!) where new ideas and innovations are safe enough to try.
Be Playful with AI. I’ve really enjoyed
’s Substack on play. His definition of play captures how I hope we learn about AI: by being creative. Ask AI to generate itineraries for travel to places you’ve always wanted to go. Use it to meal plan. Have it write things in the voices of famous authors. Tell it a joke and see if it can guess the answer. Ask it to tell you a joke. See if it can surprise you.Up until yesterday, this post was called “Build Educator Comfort and Confidence with AI.” That title was poorly chosen. We shouldn’t be aiming for comfort or confidence. AI should be making us uncomfortable, and wary, and worried. But, we as educators also have a job to do this coming school year: to talk to our students about AI and what it means for their work in and beyond school. To do that, however uncomfortable we are, we need to make sense of it for ourselves.
What would you add to this list? What are you thinking about when it comes to generative AI in education? Let’s talk more in the comments.
Links!
Powerful multimedia exploration of how generative AI tools reflect and exacerbate biases in our society.
This Blueprint for an AI Bill of Rights would be a wonderful resource for conversations about creating school policies and principles around use of AI.
I love the clarity and learning-centered focus of Australia’s draft National AI in Schools Framework.
Anyone following AI in education should be following
’s Substack “One Useful Thing.” Start with “The Homework Apocalypse.”
Thanks for reading! If you enjoyed this post, I hope you’ll share it with others. It’s free to subscribe to Learning on Purpose. If you have feedback or just want to connect, you can always reach me at eric@erichudson.co.
Thanks for this piece, Eric. You pointed out something I hadn’t considered before, which was the psychological safety of trying out new ideas with an AI partner - that feels like such a huge win.