Thinking Outside the Bot: A Conversation with Bodong Chen & Seiji Isotani

Bodong Chen: The idea was to create a space where faculty with very different perspectives on AI could think together—without an agenda, without rushing to solutions. We have technical experts, policy thinkers, historians, philosophers, people working in schools and communities. Normally, we’re all too busy to pause and ask the big questions. The Thinkery gives us that space. We’re only four meetings in, but we already have 12–15 faculty regularly participating.

Seiji Isotani: What makes it special is that it’s intentionally open. People come because they want to contribute, not because they were chosen. And when you’re not trying to accomplish something specific, creativity emerges. Innovation needs that kind of freedom. We can ask, “When is AI good? When is it not? Who benefits? Who doesn’t?” Because nobody knows everything about AI right now.

Chen: Absolutely. And that has helped in the creation of a scale—

Isotani: An AI in Education Fear Scale. We want to help people to understand if the pushback against the use of AI is based on an empirical understanding or if it’s just fear of the unknown—fear of it taking your job—so we can understand how it interferes in your use or adoption of AI.

Chen: It’s a phenomenon that we experience in different ways—the fear of AI taking over or making decisions for us as learners or as teachers. It’s pretty real, but having this additional instrument to make sense of it, I find useful.

Chen: GSE has a unique role to play as an education school. Penn is starting to build more cross-school structures, like the Penn AI Council, which Seiji sits on. There’s real interest—from the Provost’s Office, from the Center for Excellence in Teaching, Learning, and Innovation—in how GSE can help shape the University’s approach to teaching and learning with AI. We’re in the early stage, but the momentum is there.

Isotani: The University launched Penn AI Month in April, where schools across campus hosted events. GSE led several [including an AI in education symposium and a panel on mindfulness, learning, and AI]. It’s all part of building shared understanding before we make big institutional decisions.

Isotani: The big question is: How can we scale the use of AI responsibly here at Penn? Some instructors want to use AI; others are reluctant. So instead of a top-down model where administrators train faculty, we’re flipping it. We pair a GSE student with a student from another Penn school—say, Arts and Sciences or Engineering. They observe a class together, analyze what’s happening, and propose ways AI might, or might not, improve the learning experience. Then we start in instructional design, developing new ways of presenting the same content, but using AI, and then we test the impact. It’s bottom-up change, driven by students who understand both the possibilities and the pitfalls. The project has a one-year Penn grant and a two-year dean’s grant here at GSE. One focuses on breadth across schools; the other digs deeply into methods and theory.

Isotani: Students are crucial. When I arrived at GSE, I was teaching the “AI Learning Sciences and Education Policy” course, and one of my students said she wanted a space where students can interact and learn more about AI. That led to the creation of the AI in education student club, AIED@Penn. It’s housed at GSE but open to all Penn students. They’ve hosted workshops with faculty speakers—I presented on “vibe coding,” which is just coding with AI for people who don’t know how to code—and they have brought sponsors as well.

Chen: I have a meeting next week to talk about how I can support them. They represent multiple programs—though mostly our Learning Sciences and Technologies program. They participate in hackathons at Wharton and elsewhere. One team made the finals. I’d love to see them run their own hackathon someday—one grounded in educational problems, not just technical challenges.

Chen: It doesn’t feel new to me. My work has always focused on human agency—how learners and teachers make decisions, how we design environments that support inquiry and thinking. Whether it was classroom tools, collaborative learning platforms, or now AI systems, the core question hasn’t changed: How do we preserve and enhance human agency? AI adds urgency because it can make learning more efficient . . . but efficiency isn’t the only value. Learning requires friction—moments of struggle that deepen understanding.

Isotani: Yes, there is always friction between efficiency and learning. People often want AI to make everything frictionless, but learning doesn’t work that way. Take reading. Many students now just ask AI for a summary. That might get you a B on an assessment, but you lose the capacity to understand what the author is trying to say by not connecting it to your own words and ideas yourself. And the summary is not without biases. The summary will remove things you won’t even know about. That’s the challenge—how do we increase the efficiency while at the same time improving our capacity to understand things. For example, I use a five-step process when I read with AI. I skim, then I ask AI to generate visualizations. This is interesting because you are connecting with multimedia learning theory, where different perspectives actually help you to grapple with the main ideas that you are reading. Maybe next I use mind maps or slides or have AI turn the text into an audio podcast. Then, you can return to the full text with a better understanding of its basic concepts. . . . Now you can actually challenge what is in the paper. I used to spend two to four hours just reading a paper. I still spend the same amount of hours, but with much deeper interaction. You don’t save time, but you gain depth. That’s the mindset we need: AI as augmentation, not replacement.

Chen: One of my current projects, InkSpire, is exploring how AI can help instructors guide students through complex readings. It connects pedagogy, cognition, and large language models—and it reflects what we want the Thinkery to be: a place where ideas from across fields come together and “dance.”