I’ve written a few times recently on the concept of friction(less) pedagogy and learning in relation to AI. At the end of Rhetorics of the Digital Nonhumanities, I devote a chapter to pedagogical design, particularly in relation to Rittel and Webber’s wicked design problems. For a number of years, EDUCAUSE’s Horizon Report employed this concept. It identified and organized a series of problems into categories: solvable, difficult, and wicked. .

As I discuss in that chapter, from 2014-2019, “digital literacy” was identified as a “solvable problem.” This somehow remained the case even as our society was torn apart by algorithms, social media, and fake news. I don’t think it has become any more solvable in the interim. As I argued in that chapter, EDUCAUSE had no idea what “digital literacy” was/is, nor if such literacy would be of use for living in a contemporary media ecology. One would have to say it has not gone well. But as I argue there, and won’t go into now, these are wicked design problems. That said, it is understandable that EDUCAUSE, a conglomerate of EdTech companies and professionals, would conceive of design as a problem solving approach. It is what they do.

In recent decades, instructional design has taken “design thinking” on board as a methodology for achieving its ends. Design thinking informs familiar business practices such as design sprints. As such, it is not surprising that such thinking would lead to the idea of designing friction into learning experiences as a solution to the apparent frictionless quality of learning with AI.

This subject came up in a recent podcast interview with my colleague (and dean) Jeff Grabill. The interview followed on a edTech conference held on campus. It seemed that there was some consensus in the room around the notions that  “A good teacher is going to engineer friction. And a good teacher’s also going to pick students up when they fall down, give them a chance to reflect on.” Of course the question is how “to engineer friction” in the context of AI. For Grabill the answer appears to be design. As he says, he thinks “educators are really struggling about where to design friction in the educational environment.” Despite this struggle he also says that “for people like me, there’s real frustration because the answers to the AI friction problem, if you will, are in front of us. We’ve known them for a very long time. We just have to get educators to spend the time and energy to engage in them.”

This reminds me of the Horizon Report and its confidence in known problems. Working among artists, I can tell you that you won’t find many fans of “design thinking.” Design signals another contested term rather than a problem-solving approach we have shared for a very long time.

For example, do I think my university has designs on learning? I’m sure it does. In the ~15 years I’ve been here, the university has invested about $1B on STEM infrastructure and not a penny on the arts and humanities. I would certainly call that having a design on learning. Its approach to AI has been similar. When an interdisciplinary AI and society department was created, arts and humanities faculty were excluded from any conversations around its formation. And my university is hardly alone in these terms. In my view, the challenges of AI are not directly about learning at all. Instead, the challenges relate to the way universities have weaponized AI to accelerate their STEM-focused designs. That’s one version of “design” on campus.

If there is a secondary design challenge for us then it is in the way AI dismantles the notion of learning design, particularly its reliance upon outcomes. “Learning outcomes” is a term of art that goes beyond general predictability toward measurability. Learning outcomes are an effort to design limits on the potential learning experience of a course. As such, they operate on a metaphysics that assumes such matters are knowable and controllable. As “learning designer,” I master the subject. I master where the students are starting so that I can master where they will arrive at an outcome. Differences among students can be accommodated but they will arrive at the same destination. I have the power of determining what students have learned through an analysis of their products. And I can design to ensure all of that happens reliably.

For example, Grabill speaks of “engineering friction.” I would say many engineering fields (e.g. mechanical engineering) are literally about engineering friction: where you want, where you don’t, and how much. The point of engineering, or even simple machines, is to transform friction into work. So this raises the question: do we need to “work” to learn? No. But we might need to work to limit learning to outcomes. That is, when we are engineering or designing friction, we aren’t actually creating “friction” we are designing work. But when we design something as work, we design it as something that can be reproduced by AI. On the other hand if we were actually talking about friction, then AI couldn’t reproduce it because AI only operates and friction as friction does not.

AI has exposed frailties in that process, but it didn’t cause them. The problem is within instructional design and learning sciences. For example, we’ve imported the concept of design sprints. The first step in design sprints (“Understand”) requires the identification of “metrics of success” (i.e., learning outcomes). So how would a design sprint relate to learning outcomes when it is based on outcomes? It’s a self-validating activity. These activities push toward predictable outcomes that were overdetermined by the institutional context. They are an excellent example of the pseudo-participation offered to faculty at contemporary universities. However their disciplines function internally, the institutional operation of learning science and instructional design is a managed foreclosure on the potential of learning.

In short, contra Grabill, I would argue that AI reveals the intellectual frailties of instructional design and design thinking. I don’t think the “solution” lies in engineering. In fact, I’ve always asserted that engineering “solves problems” by creating future problems. Fundamentally in the humanities we confront the problem of finitude I suppose, but I’m not looking for a solution to it. Similarly I’m not interested in solving the problem that AI changes human experience. I do not imagine this problem as having a design solution.

And efforts to bring learning design to the challenges AI presents to higher education are sure to become another site of conflict. In my view, it represents another effort to impose some misplaced hierarchical authority over academic life.

.