Faculty Are Redesigning Alone. Here Is the Infrastructure That Changes That.
The Chronicle of Higher Education surveyed more than 100 faculty members this spring about how generative AI has reshaped their teaching. The publication also drew on a California State University survey of 94,000 students, staff, and faculty members. What emerged is less a picture of AI disruption than a picture of what happens when faculty are handed a systems-level problem to solve on their own.
The numbers split into two groups. On one side, faculty are playing defense: more than 80% have tried to AI-proof assignments and assessments, and 65% have caught students using AI to cheat. On the other side, faculty are adapting: nearly half have introduced assignments where students work with AI, and 55% of Cal State faculty use AI to develop course materials.
What the data doesn't show is where those faculty are getting their design frameworks. The answer, across interview after interview in the Chronicle piece, is that they're building them from scratch.
What It Means for Design
Every faculty member in this piece is doing something genuinely hard. A philosophy professor at San Jose State built a low-tech pedagogy layer, distributing paper course packs, requiring in-class participation, and using shared Google docs to trace student thinking through the writing process. A computational chemistry professor instituted oral exams modeled on industry coding interviews. A psychology professor built a three-page syllabus section on AI policy alongside a multi-step paper verification system.
These are good design instincts. They are also the product of individual faculty members solving a shared institutional problem: without support, without a framework, and often without time to sustain the solution past a single semester.
Faculty have ideas. The scale of what they're doing individually signals a design infrastructure gap at the institutional level. Distributed ingenuity, however genuine, doesn't accumulate into a system. The Chronicle quotes a history professor whose hallway conversations have shifted entirely to AI policing. An English professor retired early, at 61, because she had run out of capacity to rebuild her courses again after Covid. These are not outliers. They represent the cost of solving an institutional problem one course at a time.
The question that belongs at the institutional level is specific: how do institutions build learning experiences where understanding has to be demonstrated, not just submitted? Right now, that question is being answered by individual faculty members, in real time, without a shared architecture beneath them.
The SRM Read
What distinguishes institutions moving forward on AI integration is a shift in the underlying design question. Policy-level responses address the symptom. Course architecture that builds authentic demonstration of learning into its structure, through oral defense, process documentation, and scaffolded iteration, changes the assessment problem entirely. The assessable thing is the thinking. Whether a tool helped produce a draft becomes secondary.
That shift doesn't happen at the policy level. It happens at the course design level, which is where SRM works.
Explore SRM's AI + Human Insight Approach