Georgia Northwestern Technical College

AI accelerated a reckoning we were already overdue for. Workforce-driven students have always been skeptical of busywork. They have no patience for assignments that don't connect to a real job and they never did. When AI tools made it trivially easy to complete low-stakes tasks without engaging with the material, they didn't create that problem. They made it visible and that visibility has been clarifying.

My response has been to help faculty shift toward what I call process-product design: assessments where the process of completing the work is itself evidence of learning, not just the final deliverable. A completed essay proves nothing on its own. What we're actually after is a student who can talk through their reasoning on the spot and apply something in a context they haven't seen before. The deliverable is the least interesting part of the evidence.

When the Deliverable Is the Least Interesting Part

Instructional design's job is to make concrete what we actually mean by learning. In a technical college context, that means working with faculty to distinguish between thin learning and the real thing. The distinction we keep coming back to is this: can this student's knowledge hold up when something goes wrong? A medical coding student who uses an AI tool to generate a lookup has demonstrated something, but not competency. The student who matters is the one who understands why a code matters and catches an error when it appears. An employer figures that out in the first week, even if the transcript never does. Instructional design helps faculty ask the question that makes this concrete: would an employer recognize this as evidence of competence? If not, the assessment needs work, regardless of whether AI is anywhere near it.

“AI tools made it trivially easy to complete low-stakes tasks without engaging with the material. They didn't create that problem. They made it visible.”

The biggest shift in faculty support over the last two years has been AI anxiety. Faculties are worried about academic integrity and about their own relevance and many are operating without a clear framework for thinking about what AI actually changes versus what it doesn't. My approach has been to move the conversation away from "how do we stop AI" and toward "what do we want students to be able to do and does this assignment actually measure that?" That reframe does something useful on both counts: it reduces the panic and it surfaces the underlying design problem, which is usually that the assignment was vulnerable long before AI showed up. Chasing detection is a losing race because the assignment itself is usually the problem.

What Labs Are Actually For

Some things genuinely cannot scale online. A student learning to take blood pressure or wire a circuit needs hands-on practice and building hybrid lab components into an online program isn't a failure of imagination” it's an honest assessment of what learning requires. The scalability argument makes more sense for the foundational knowledge that supports hands-on work: the reasoning and safety protocols a student needs before they ever pick up a tool. Moving that content online frees up lab time for what labs are actually for. The discipline is making that distinction deliberately, rather than defaulting to "everything can go online" or refusing to try.

Authentic Assessment Will Look Smart Soon

Two things are worth watching and they're connected. The first is AI's continued pressure on assessment design. Autonomous AI agents can now complete entire online courses with minimal human involvement that's documented, it's spreading and it will force a long-overdue reckoning with what we're actually measuring in online learning.

The second is what I'd call the authenticity turn: a growing emphasis on assessments that require personal, situated knowledge, which are inherently harder to shortcut. An assignment built around a student's specific workplace, or a portfolio that accumulates over a program rather than a single semester, can't be dispatched the same way a generic prompt can. For workforce education, this is a natural fit. The field has always been oriented toward outcomes that matter outside the classroom and institutions that lean into authentic, performance-based design now are going to look very smart in five years. The alternative is spending those years fighting a battle that gets harder every semester.