A reaction to Educause Exchange’s conversation on generative AI in teaching and learning
Recently, I started spending more and more of my time in the compliance and accessibility corner of higher-ed, so I watch these AI-in-the-classroom conversations as a bona-fide outsider. I’ll likely never have to grade an essay or design a course. Still, I find myself curating the range of opinions floating around, trying to figure out which ones will still be true in two years.
What struck me watching this particular conversation wasn’t any single insight. It was how many different people, coming from different corners of the institution, kept landing on the same word without seeming to coordinate on it: belonging.
The part that isn’t really about AI
Most of the AI-in-education conversation gets framed as a technical or policy problem. Detection tools, honor codes, rewritten syllabi. This conversation kept drifting somewhere else instead. One person talked about students needing to maintain their own voice and perspective. Another talked about creating a safe space where faculty and students can share what they’re actually worried about, not just what they’re supposed to say in a training session. Someone else described the whole moment as something faculty and students are learning together, as peers, rather than one group managing the other.
That’s a belonging problem before it’s an AI problem. The anxiety under a lot of this conversation isn’t really “will students cheat.” It’s “will students and faculty still recognize each other, and themselves, once a machine can produce a passable version of almost anything either of them makes.” A professor’s expertise used to be legible through the artifacts they produced and assigned. A student’s growth used to be legible the same way. When the artifact stops proving anything, the thing it was standing in for, a sense of who belongs in this work and why, has to get rebuilt on different grounds.
I don’t think anyone in this conversation had that fully worked out. What I liked was that nobody pretended to. There was real range here: someone confident that reflection with a chatbot unlocks something genuinely new, someone more cautious about AI literacy and people trusting outputs too much, someone thinking at the scale of accreditation and whether large public institutions can even move at the pace this requires. Curating that range felt more honest than picking a side.
Where I think this actually lands
If I had to bet on the piece of this that survives the hype cycle, it’s the shift toward process over product. Not because it’s a clever pedagogical trick, but because it’s the only framing that keeps belonging intact. If what gets graded is the artifact, AI flattens everyone into the same output and belonging erodes, since anyone can generate the same essay. If what gets graded is the process, the judgment calls, the reflection, the documented back-and-forth with a tool, the individual is still visible in the work. That’s not a minor implementation detail. It’s the whole ballgame.
Where YuJa fits into this
This is the part I actually work on, so I’ll be direct about it rather than pretending it’s incidental.
A lot of the labor that stands between an institution and this better version of things isn’t pedagogical at all. It’s mechanical. Making sure a semester of lecture video has audio descriptions. Making sure a decade of PDFs are structurally tagged for a screen reader. Checking that captions exist across a catalog nobody has fully audited in years. None of that work builds belonging or tests judgment. It’s overhead, and it’s the kind of overhead that eats the time and attention faculty need for the actual redesign work this moment calls for.
That’s the lane YuJa’s AI tools sit in. YuJa Panorama andYuJa Lumina’s GenAI Video PowerPack use AI to clear that backlog: automated audio descriptions, caption checks, document remediation, video chaptering. YuJa EqualGround andYuJa CivicGuard pair AI remediation with human review rather than assuming full automation, because compliance work is exactly the kind of domain where getting it wrong has real consequences for someone who depends on it. None of it replaces the judgment a faculty member brings to redesigning an assignment around reflection instead of output. It just tries to make sure that faculty members have the time to do it.
I don’t think AI in higher-ed is a solved problem, and I’m skeptical of anyone who talks like it is. But watching this conversation, and thinking about where the tools I work with actually fit into it, I keep coming back to the same read: the technology that matters most right now isn’t the technology doing the thinking. It’s the technology quietly getting the mechanical work out of the way so the humans have room to do theirs.
YuJa.com
Contact Us
