On campus, students are asking practical and philosophical questions about learning and life in the age of AI. Here’s how to host a conversation on your campus.
Hello everyone, I’m a professor at Parsons School of Design where I teach Economics & Ethics of Sustainable Design and Global Professional Practices. Over the past few years, I’ve been integrating generative AI into my teaching in a way that helps students think more critically, not less. My classroom uses a full ecosystem of tools, ChatGPT, market data platforms, scenario modeling, design thinking frameworks, to help students understand how emerging technologies shape global systems, creativity, sustainability, and professional practice.
I joined this forum to learn from other educators experimenting at the intersection of pedagogy and AI. I’m also excited to share what’s been working in my courses, particularly around reflective learning, student agency, and building contextual reasoning skills with AI.
Looking forward to connecting and contributing to this community.
I like your viewpoint on learning with AI. When the question of students using AI to cheat as opposed the using AI as a value added tool in learning arises, how do you address these concerns?
Thanks, Todd. Yes,it’s not complex, but it is definitely complicated, and I’ve been following it closely since the very first release of ChatGPT. In the beginning, of course, students everywhere used it to do most of the work for them. It was new, it was convenient, and it felt like a kind of magic. I remember seeing the Microsoft CEO talk about how he learned math with it, and stories about people in remote villages in India, suddenly being able to navigate government forms in their own language. That early moment made it clear that the technology wasn’t just a gadget. It was an equalizer and here to stay.
Since then, both the academic community and the technology itself have evolved. ChatGPT now has a recognizable “signature style” - - - the dashes, the commas, the connective phrasing, and you can tell when a student hasn’t touched or shaped the output. The novelty has worn off, and what we’re left with is the practical question of how to integrate it into learning in a meaningful way. One thing that really stands out for me, especially after teaching predominantly international students for decades, is how overnight, writing and grammar improved. But more importantly, students could finally move fluidly between languages. They could understand, process, and express ideas that previously felt out of reach because of language barriers. For many of them, that was transformative.
I emphasize to them that they must read the entire output carefully and revise it until it becomes their own thinking. I tell them to use it the way you would use a calculator for long division: not to avoid the work, but to shape the outcome you need through engaged, iterative interaction. When they do that, every student’s process, and every result, becomes distinct.
Again, this technology isn’t going away, it's here to stay for the foreseeable future. So, our responsibility is to figure out how to use it well and ethically. Honestly, I sometimes see what we call “cheating” as a form of human creativity in its earliest, least refined form, trying to solve a problem with the tools available. But the real value is when students move past that and use AI as a catalyst for deeper engagement, clearer thinking, and better outcomes. That’s where the learning happens. And that’s where the tool becomes empowering rather than substitutive.
Thanks for the deep response. I agree and I've been suggesting in other forums that we need to embrace the technology and teach the learners how to best use it. I also believe that it is changing the entire learning landscape.
How do you feel about moving away from the traditional industrial classroom and into a more personalized learning approach with AI acting as a tutor?
Yes, I think moving beyond the traditional industrial classroom is not only possible but necessary. It actually brings to mind John Taylor Gatto’s critique: that our schooling system was built during the Industrial Revolution to produce standardized workers, not independent thinkers. Today that logic simply doesn’t hold. Learners anywhere, in any geography or language, can now access knowledge piecemeal, on demand, exactly when they need it. That’s a revolutionary shift. As LLMs gain more reasoning and memory, the need for “edu-influencers” selling secret formulas fades too; why buy a script when you can generate your own, tailored to your context. This changes my role as a teacher. More and more, I see myself as a guide or coach. There are moments when I can’t “teach” in the old sense, but they can learn. My job is to create an environment where students recognize their own “aha” moments, and to use a Socratic approach to help them discover meaning rather than receive it. AI tutors make that personalization possible. They free the quest from the industrial model and allow us (I hope… paradoxically) to build truly human learning experiences, where the teacher becomes the one who cultivates insight, not the one who delivers content.
Thanks again for another deep thought answer. How do you think we can effect a move in this direction? I immediately go to personalized learning plans. Imagine this: a small child begins with a plan which includes basic math, reading and writing. Then through observation, the plan starts taking a learning path which is in the best interest of the individual learner. By the time they enter a level of post secondary education, they're now on a path for adult specific, and topic specific learning. What are your thoughts?
Now you’re moving us from the complicated into complexity, and I mean that in the best way. Earlier we were talking about systems within the Industrial model; now we’re touching the deeper philosophical currents that have shaped learning for millennia. I’ve taught design and management for years, and I often remind students that “Design” has no single definition. It has existed across civilizations under different titles, embedded in different knowledge systems, each shaped by its own inputs, outputs, and interpretations. When we draw connections from the Socratic method to Ibn Sina (Avicenna), Averroes of Cordoba, or al-Farabi from Turkistan... we realize that our modern educational debates are inheriting centuries of epistemology that traveled through time and translation.
Bringing this back to AI and personalized learning: your example of a small child beginning with basic math, reading, and writing is a good entry point, but once we move beyond linear progressions into real human development, we hit questions that Chomsky raised about innate linguistic structures versus the older idea of tabula rasa. A child’s path isn’t simply a curriculum arc, it’s a dynamic interplay between absorption, flexibility, and the eventual fossilization of habits and cognition. Those developmental epochs, infancy, early childhood, late childhood, teens, twenties, all shape how learning unfolds.
This is why I see my role increasingly as a guide or coach, not a content-deliverer. There are moments when I cannot “teach,” but the student can still learn. My work becomes setting up the environment, using a Socratic approach, and helping them find the “aha” moment that arises from within, not from me.
Your question (“How do we move toward individualized learning paths?”) is important, but we must recognize the limits of current AI. I often tell students: imagine standing on a street corner looking at a new university building; now shift your position to the opposite corner and look again. That act of embodied perspective-taking is something LLMs cannot do. They operate through coded pattern recognition, not lived consciousness. So, guiding a child’s trajectory still requires a conscious human being who can recognize temperament, strengths, and emerging interests, and also be aware of their own conditioning, which becomes an input into the child’s development. AI cannot yet replace that (and this opens entire new sets of questions, problems, and challenges... anyway). Your vision is absolutely worthy. The real challenge is not simply the idea, but how we shape the path so that the outcome serves the child rather than the system, the market, or the biases of the code.
That is the question we now face: How do we design personalized learning in a way that respects human complexity, avoids overcoding the child, and still uses AI as a tool rather than a director? I believe the next phase of these human conversations must be about process, not just vision, and about who shapes that process... I wonder what Huxley would say... :)
I like the idea of using a physical book and talking to students one-on-one about their AI use. I’ve also seen that students are eager to talk about what they’re doing with each other and with their teachers.
Hi — I published a public open letter to OpenAI that went out on a national press wire a few weeks ago in the US and UK.
It documents a long-horizon creative project using a proprietary OS I built inside of ChatGPT. The AI works strictly as a manager and memory system, not to generate the art.
Hello everyone, I’m a professor at Parsons School of Design where I teach Economics & Ethics of Sustainable Design and Global Professional Practices. Over the past few years, I’ve been integrating generative AI into my teaching in a way that helps students think more critically, not less. My classroom uses a full ecosystem of tools, ChatGPT, market data platforms, scenario modeling, design thinking frameworks, to help students understand how emerging technologies shape global systems, creativity, sustainability, and professional practice.
I joined this forum to learn from other educators experimenting at the intersection of pedagogy and AI. I’m also excited to share what’s been working in my courses, particularly around reflective learning, student agency, and building contextual reasoning skills with AI.
Looking forward to connecting and contributing to this community.
I like your viewpoint on learning with AI. When the question of students using AI to cheat as opposed the using AI as a value added tool in learning arises, how do you address these concerns?
Thanks, Todd. Yes,it’s not complex, but it is definitely complicated, and I’ve been following it closely since the very first release of ChatGPT. In the beginning, of course, students everywhere used it to do most of the work for them. It was new, it was convenient, and it felt like a kind of magic. I remember seeing the Microsoft CEO talk about how he learned math with it, and stories about people in remote villages in India, suddenly being able to navigate government forms in their own language. That early moment made it clear that the technology wasn’t just a gadget. It was an equalizer and here to stay.
Since then, both the academic community and the technology itself have evolved. ChatGPT now has a recognizable “signature style” - - - the dashes, the commas, the connective phrasing, and you can tell when a student hasn’t touched or shaped the output. The novelty has worn off, and what we’re left with is the practical question of how to integrate it into learning in a meaningful way. One thing that really stands out for me, especially after teaching predominantly international students for decades, is how overnight, writing and grammar improved. But more importantly, students could finally move fluidly between languages. They could understand, process, and express ideas that previously felt out of reach because of language barriers. For many of them, that was transformative.
I emphasize to them that they must read the entire output carefully and revise it until it becomes their own thinking. I tell them to use it the way you would use a calculator for long division: not to avoid the work, but to shape the outcome you need through engaged, iterative interaction. When they do that, every student’s process, and every result, becomes distinct.
Again, this technology isn’t going away, it's here to stay for the foreseeable future. So, our responsibility is to figure out how to use it well and ethically. Honestly, I sometimes see what we call “cheating” as a form of human creativity in its earliest, least refined form, trying to solve a problem with the tools available. But the real value is when students move past that and use AI as a catalyst for deeper engagement, clearer thinking, and better outcomes. That’s where the learning happens. And that’s where the tool becomes empowering rather than substitutive.
Thanks for the deep response. I agree and I've been suggesting in other forums that we need to embrace the technology and teach the learners how to best use it. I also believe that it is changing the entire learning landscape.
How do you feel about moving away from the traditional industrial classroom and into a more personalized learning approach with AI acting as a tutor?
Yes, I think moving beyond the traditional industrial classroom is not only possible but necessary. It actually brings to mind John Taylor Gatto’s critique: that our schooling system was built during the Industrial Revolution to produce standardized workers, not independent thinkers. Today that logic simply doesn’t hold. Learners anywhere, in any geography or language, can now access knowledge piecemeal, on demand, exactly when they need it. That’s a revolutionary shift. As LLMs gain more reasoning and memory, the need for “edu-influencers” selling secret formulas fades too; why buy a script when you can generate your own, tailored to your context. This changes my role as a teacher. More and more, I see myself as a guide or coach. There are moments when I can’t “teach” in the old sense, but they can learn. My job is to create an environment where students recognize their own “aha” moments, and to use a Socratic approach to help them discover meaning rather than receive it. AI tutors make that personalization possible. They free the quest from the industrial model and allow us (I hope… paradoxically) to build truly human learning experiences, where the teacher becomes the one who cultivates insight, not the one who delivers content.
Thanks again for another deep thought answer. How do you think we can effect a move in this direction? I immediately go to personalized learning plans. Imagine this: a small child begins with a plan which includes basic math, reading and writing. Then through observation, the plan starts taking a learning path which is in the best interest of the individual learner. By the time they enter a level of post secondary education, they're now on a path for adult specific, and topic specific learning. What are your thoughts?
Now you’re moving us from the complicated into complexity, and I mean that in the best way. Earlier we were talking about systems within the Industrial model; now we’re touching the deeper philosophical currents that have shaped learning for millennia. I’ve taught design and management for years, and I often remind students that “Design” has no single definition. It has existed across civilizations under different titles, embedded in different knowledge systems, each shaped by its own inputs, outputs, and interpretations. When we draw connections from the Socratic method to Ibn Sina (Avicenna), Averroes of Cordoba, or al-Farabi from Turkistan... we realize that our modern educational debates are inheriting centuries of epistemology that traveled through time and translation.
Bringing this back to AI and personalized learning: your example of a small child beginning with basic math, reading, and writing is a good entry point, but once we move beyond linear progressions into real human development, we hit questions that Chomsky raised about innate linguistic structures versus the older idea of tabula rasa. A child’s path isn’t simply a curriculum arc, it’s a dynamic interplay between absorption, flexibility, and the eventual fossilization of habits and cognition. Those developmental epochs, infancy, early childhood, late childhood, teens, twenties, all shape how learning unfolds.
This is why I see my role increasingly as a guide or coach, not a content-deliverer. There are moments when I cannot “teach,” but the student can still learn. My work becomes setting up the environment, using a Socratic approach, and helping them find the “aha” moment that arises from within, not from me.
Your question (“How do we move toward individualized learning paths?”) is important, but we must recognize the limits of current AI. I often tell students: imagine standing on a street corner looking at a new university building; now shift your position to the opposite corner and look again. That act of embodied perspective-taking is something LLMs cannot do. They operate through coded pattern recognition, not lived consciousness. So, guiding a child’s trajectory still requires a conscious human being who can recognize temperament, strengths, and emerging interests, and also be aware of their own conditioning, which becomes an input into the child’s development. AI cannot yet replace that (and this opens entire new sets of questions, problems, and challenges... anyway). Your vision is absolutely worthy. The real challenge is not simply the idea, but how we shape the path so that the outcome serves the child rather than the system, the market, or the biases of the code.
That is the question we now face: How do we design personalized learning in a way that respects human complexity, avoids overcoding the child, and still uses AI as a tool rather than a director? I believe the next phase of these human conversations must be about process, not just vision, and about who shapes that process... I wonder what Huxley would say... :)
An amazing experience, can't wait for more students and educators to join the conversation!
Thanks for leading the way at Penn!
I have been writing about this for a year and a half yes.
Woooo go Katie
This is an important discussion. So glad you are doing it.
Hi Team,
Do organizations outside the USA or Canada eligible to apply to some of these programs or opportunities?
Hey Herman, we'd be happy to send you our guide and digital materials, we just won't be able to send physical copies of the book. Go ahead and fill out the form: https://docs.google.com/forms/d/e/1FAIpQLSf-51q8VYGMMkEVsaklikW-NdZXPHIFFuGji0eg2gTP26L6jA/viewform?usp=dialog
Katie's at OpenAI!?! Heck yes :)
Hello!
Only for students in USA?
I want to use in highscool in Romania, form my kid, he is 18 years old, in last year before college...
Hey Radu, we'd be happy to send you our guide and digital materials, we just won't be able to send physical copies of the book. Go ahead and fill out the form: https://docs.google.com/forms/d/e/1FAIpQLSf-51q8VYGMMkEVsaklikW-NdZXPHIFFuGji0eg2gTP26L6jA/viewform?usp=dialog
I like the idea of using a physical book and talking to students one-on-one about their AI use. I’ve also seen that students are eager to talk about what they’re doing with each other and with their teachers.
It is great to get these insights. Is their an objective to these engagements?
Hi — I published a public open letter to OpenAI that went out on a national press wire a few weeks ago in the US and UK.
It documents a long-horizon creative project using a proprietary OS I built inside of ChatGPT. The AI works strictly as a manager and memory system, not to generate the art.
Sharing in case it’s of interest.
https://open.substack.com/pub/scooterscudieri/p/open-letter-to-open-ai-post-14?r=1iz56i&utm_medium=ios
https://open.substack.com/pub/teodoracoach/p/the-1-billion-secret-what-the-fastest?utm_source=share&utm_medium=android&r=6zsoq4
I piloted front-facing bespoke AI modules in my class last semester and I’ll be presenting my research at the ICEDU pedagogy conference this March!
Thank you so much. My pinned post has thoughts on creativity and provenance, food for thought for teachers and students alike
Hi Katie,
Thanks. We shall be happy to receive the material.
AI helps us immensely. But we are still beings with brains from a prehistoric age.