I've been trying to help people learn things for most of my life. (Ask me about "Infozine" sometime.) I minored in English Language Arts education in college, spent time in a public elementary school as a tutor and teaching assistant, and ran after-school programs and summer camps for school-age kids before I ended up in marketing. As a marketer, I've spent over twelve years explaining complex products to people who didn't know they needed them. It's teaching in a different context. The rules and techniques are different, but they're both built on psychology and communication.
Recently I began noticing that smart, experienced friends and colleagues were reluctant to adopt AI. Often they used it like a slightly better search engine. Sometimes they would ask it to do meaningful work. The results were generic and often wrong, and they'd walk away more skeptical than before. They assumed AI was inadequate or that they didn't have the technical expertise to use it properly. I had a different take: they didn't have an effective mental model for how to work with it.
AI can approximate general expertise in almost any domain. What it can't replicate is what you know from years inside a specific organization, working on specific products, where the most important insights are embedded in an ambiguous context that only someone with firsthand experience can parse. When I understood that, everything changed. I stopped asking AI to do my job and started treating it as a thinking companion (thank you, Ethan Mollick). The results were different enough that I wanted other people to have the same experience.
AI Apprenticeship is what I built in response. The framework is a five-stage structure for working through real projects with AI, developed across four applications I built without a technical background. Hora is the application that guides users through it. None of this was built to demonstrate what AI can do. It was built to work out, in practice, how a non-technical person develops real capability with AI by doing real work.
My daughters are learning to read right now, and I've spent more time than I expected thinking about how that actually happens: not through exposure alone, but through structure, repetition, and the gradual removal of scaffolding as capability develops. That's what the AI Apprenticeship framework is meant to do.
What interests me about AI, beyond what it has already made possible, is what it could do for learning specifically. Through AI, we now have the ability to give every learner a patient, knowledgeable guide who responds to where they actually are, not where the curriculum assumes they should be. That possibility hasn't been fully realized yet. Working out how it gets realized is where I want to spend my time.
If any of this is relevant to what you're working on, I'd like to hear about it. You can reach me at pmrotter333@gmail.com.
- Paul