A Year Ago, I Panicked. (Again.)
A year ago, I was deeply uneasy about the sheer ubiquity of the Tech-Whose-Name-We-All-Dread (TWNWAD). It was everywhere, dominating the news, colonizing our collective headspace, and creeping, rather insidiously, into my classroom.
Suddenly, students were submitting these beautiful looking, very mediocre assignments. Impeccable grammar. Endless em dashes. Polished prose. And the most painfully trite arguments imaginable, all delivered in response to carefully designed written prompts.
This, of course, triggered yet another existential crisis. (I get those when confronted with particularly bad papers.)
The horrifying thought: everything I had learned, everything I had taught so far, was slowly becoming useless. Obsolete.
And then came the second realization, arguably worse. TWNWAD was here. It wasn’t going away. And pretending it didn’t exist was no longer an option.
Enter the Sane Saviors
Thank the universe for sane, thoughtful guides, most notably Ethan Mollick, who operate in the same management education space but talk not about moral panic, but about embracing, learning, and evolving alongside TWNWAD. His book Co-Intelligence was, quite genuinely, a lifesaver.
So I asked myself:
How do I understand this thing and adapt it for the classroom so that it becomes a pedagogical tool rather than an academic menace?
That question eventually became AI in Action.
How AI in Action Was Born
AI in Action is a semester long project I designed for my graduate course, MGMT 628: Human Resource Development, first introduced in Winter 2026. The core idea is simple. If AI is reshaping Learning and Development, then students shouldn’t just talk about it. They should build with it.
In this project, students learn about artificial intelligence by creating their own AI powered tools. Specifically, they build a Retrieval Augmented Generation (RAG) training bot designed to deliver HRD content.
The project unfolds in stages:
• Students author professional training documents on a chosen HRD topic
• They design the bot’s persona and system prompt
• They upload and curate content.
• They test the bot extensively.
• Finally, they produce an analytical report examining bias, data security, hallucination risks, trust, and regulatory compliance.
Students critically engage with bias amplification, data security, hallucinations, and ethical risk. They grapple with transparency and accountability in AI driven training systems and explore how regulation plays out in corporate Learning and Development contexts. These ideas surface in the term paper and classroom debates, so the technical and societal dimensions of AI stay tightly linked.
In designing the project, I followed AACSB’s guidance for teaching AI across three key domains: understanding the technology, using the tools, and evaluating their impact.
And honestly, Learning and Development is the perfect test bed. SHRM keeps reminding us that AI is rapidly reshaping L&D, so where else should we experiment?
Building the Bot (and Loving My Students even more)
Armed with an EMU e Fellows grant (which covered a Claude Pro license) and more than a little optimism, I set out this semester.
The results exceeded expectations.
My students were skeptical, which I loved. Overly compliant people don’t question assumptions, and they certainly don’t help with exploration. We started by developing training materials on metrics used in Learning and Development. Our textbooks mention these, but never consistently. Worst case scenario, we’d at least walk away with a solid glossary of T&D metrics.
Then we defined the learner persona and the training bot persona, because, training without a persona is just depressing.
Finally, we built a no code RAG bot using Claude Pro.
Click here to access our Training Bot
There are moments when I absolutely adore my students. Our bot turned out to be mean. Unkind. Sarcastic. Possibly a little too much. But mercifully, it wasn’t the cloying, aggressively encouraging creature we initially imagined.
And somehow, it worked. Really well.
What This Project Actually Did
This project wasn’t just about wrestling with a new technology. Sure, we could have coded more elegantly. We could have reduced bias further. We could have refined the language.
But it forced me to think deeply about what I teach and why. I found myself becoming philosophical about the nature of work and what parts of it are genuinely human versus purely mechanical.
Designing this experience pushed me to reflect on training design, content creation, evaluation, and most crucially, the learner experience itself. I learned far more about Learning and Development by trying to teach it through AI than I ever expected.
Taking It Public
At the Gen AI Spring Summit 2026 at Eastern Michigan University, I had the privilege of sharing this journey during a panel discussion alongside one of my students. Reflecting together on how the project challenged both our technical abilities and our professional assumptions was deeply rewarding.
The audience response was phenomenal.
Yes, we had the usual code checkers telling us all the ways we were “wrong” (useful, actually). But we also had incredibly thoughtful conversations about what the bot was, what it did, and how its architecture shaped the interaction itself. Watching people engage so seriously, and curiously, with the project reminded me why I do this work.
Because sometimes, confronting the TWNWAD doesn’t mean surrendering to it.
It means learning how to argue back, thoughtfully, critically, and maybe with a little sarcasm. And always the coffee!
No comments:
Post a Comment