I have been teaching a version of Education by Design for longer than I can remember. I designed it when I was at Michigan State and carried it over when I came to ASU. And to say that I love teaching this class would be an understatement. It brings together everything I care about: education, design, creativity, and whatever technology is shifting the ground under us. This is why, though the underlying themes and readings have stayed pretty steady it has never been the same course twice. That is partly due to me and my temperament, to never do the same thing twice. But it is also that a course about design should keep answering to the world it is designing for, and that world keeps moving.
For the last few years the thing moving fastest has been generative AI, and readers of this blog know that I have no faith in either of the standard institutional responses, prohibition or wholesale uncritical adoption. Both these responses rest on a shared assumption that we can settle what this technology means for learning before we teach with it. We cannot, and one reason we cannot is that AI was never designed for educational purposes in the first place. It arrived from somewhere else, built for other reasons and other markets, and the work of figuring out what it is good for in a classroom falls to the people standing in the classroom.
That is a design problem in the technical sense of the term. You cannot fully specify it before you start working on it, every attempt at a solution changes the problem, and no test tells you when you are finished. Design is the discipline that has always operated under those conditions. So the fall 2025 version of the course ran on the Five Spaces of Design in Education, which asks students to work across artifacts, processes, experiences, systems, and culture as distinct but connected places where design happens. And the engine that drove this enterprise was Dewey’s four primary impulses for learning: inquiry, communication, construction, and expression. So there were no restrictions on GenAI use and students were asked to treat it as a creative partner and to disclose and cite what it had done for them. As the syllabus stated “standard rules of academic integrity still apply”—do not take credit for something created by someone (or something) else.
This iteration of Ed/Design was unique in terms of the students in the room. I make this seminar open to all majors and advertise it heavily through my networks. My experience has taught me that this class is stronger when I have a diverse group of participants, and this semester was the most diverse ever. We ended up with twenty-five participants, including high school students, master’s students, doctoral students, and university faculty, all doing the same assignments and all treated as peers. I did not fully anticipate what that would do. When nobody in the room can plausibly claim authority over a technology this new, the usual hierarchy stops being useful, and what replaces it is a real distribution of expertise. One student described it this way: “There was no age limit, no strict divide between backgrounds, no sense of hierarchy… that diversity made discussions richer and made me learn more from my classmates than I expected.”
The work ran at three scales (micro, meso and macro), and in each case the assignment fixed a form and left the content wide open. The micro assignments focused on weekly readings and reflections, a running “Good and Bad Designs” series where students shared examples from their own lives, and a rotating documentarian who had to turn each session into a creative artifact of their own choosing, which over the semester produced podcasts, poems, videos, infographics, and interactive mind maps. The meso projects asked for 45-second radio commercials for fictitious AI companies, video reels introducing design theorists, and a two-week response to our Vice Provost’s challenge to reimagine education for populations the system has historically excluded. The (macro) individual semester project named no topic at all, only that students find an educational design problem they cared about and build something that worked, to showcase at an end of semester AI Carnival. Students created everything from historical museum of technology to a carbon calculator, a music production tool, and a web builder.
None of this produced settled answers about GenAI and education, and I did not expect it to. What the students left with, I hoped, and in many cases was reflected in their final reflections, was the capacity to inquire, construct, communicate, and express in conditions where nobody knows what comes next. Uncertainty was the condition of the work rather than an obstacle to it, and design was what let us work inside it.
One student wrote in their final paper that “design and anxiety are siblings. Both live in the space between what exists and what could be. Anxiety means you care, it means you’re paying attention. It pushes you to question, to refine, and to imagine again.” She went on to write about what getting there had cost: “Unlearning, I’ve discovered, feels like standing barefoot on a cold floor after you’ve been wearing shoes for years. You’re suddenly aware of every texture, every edge, every discomfort.”
The full account is now out in the Journal of Instructional Design and Technology, written with Emmanuel Adeloju (who was a student in the class), along with Lindsey McCaleb and Nicole Oster. It is open access.
Adeloju, E., Mishra, P., McCaleb, L., & Oster, N. (2026). Designing in Uncertainty: GenAI and Creative Agency in a Graduate Seminar. Journal of Instructional Design and Technology, 1(2), 1–7. https://doi.org/10.65201/RKHM9093
Incidentally, I am teaching the next iteration of the course this semester – and yes the readings may be the same but the projects are very different. More on that later.