When I read Carlo Rotella’s recent essay (I’m a Professor. A.I. Has Changed My Classroom, but Not for the Worse) about teaching in the age of AI, one student line lodged in my mind and wouldn’t let go:
How can we “ask the questions on our own” when there isn’t a teacher there to do it for us?
That question, posed by a student named Tyler at the end of an English class at Boston College, is the real heart of Rotella’s piece. It’s also the heart of what I’m trying to build with Topoi.
Rotella describes his classroom as a workshop, not a factory. Students don’t sit passively while content is delivered; they practice a way of coming at things—noticing patterns, asking better questions, building interpretations, arguing for what they see in the text and in the world.
If you’ve spent time in humanities classrooms, you’ll recognize the scene: marked-up paperbacks, messy notes, blue books, conversation that stutters and then catches, the slow construction of shared meaning. Rotella’s point is simple and powerful: in an era of generative AI, the best defense of that kind of education is not more technology, but more humanity—more attention to process, to community, to the actual work of thinking.
Topoi agrees with him completely.
And then it asks: what if our tools could be redesigned to support exactly that kind of work?
AI-Resistant vs. AI-Oblivious
Rotella is not in denial about AI. He knows students can outsource reading and writing to a bot. He knows they sometimes do. But instead of treating AI as an enemy to be detected and policed, he treats it as a pressure that clarifies what matters.
His response has three main moves:
- Assessment that foregrounds human presence.
Pen-and-paper quizzes, blue-book exams, oral work—tasks where students have to show up as themselves. - Teaching writing as a process, not just a product.
Scaffolding assignments, drafts, conferences, peer review; paying attention to how the paper comes into being, not just the final PDF. - Doubling down on the classroom as a community.
Device-free rooms, everyone speaking, everyone known by name. Students are not buying content; they are buying access to each other and to a shared practice of thinking.
This is what he calls an “AI-resistant” course. Not AI-proof—there’s no such thing—but deliberately structured so that the value proposition is human.
Topoi lives in the same neighborhood. It starts from the same convictions:
- Reading is thinking. Writing is thinking.
- Process is where learning happens.
- Community is the irreplaceable context for deep work.
Where Topoi differs is not in values but in architecture.
Topoi: Building Tools That Respect the Workshop
If you’ve only experienced AI as a one-on-one chatbot in a browser, it’s easy to see it as the thing Rotella rightly resists: a way to skip the slow, effortful parts of learning.
Topoi starts from another question:
If the classroom is a workshop, what kind of AI infrastructure would actually serve that workshop?
Concretely, that leads to some design commitments that line up closely with the world Rotella wants:
- Human-initiated only.
The Star Trek computer doesn’t interrupt the bridge; Topoi doesn’t interrupt a class. AI is called on by humans when needed, not constantly pushing suggestions or answers. - Group, not solo, by default.
Rotella’s classroom is a place where students “see and hear one another thinking.” Topoi is built around groups—teachers, students, peers—working in shared sessions, not isolated chats. AI is a resource the group consults, not a private accomplice. - Process is visible.
Rotella reads drafts, margins, and mark-ups because that’s where thinking lives. Topoi preserves transcripts, turns, prompts, and revisions so that teachers and students can see how an idea evolved, not just what the AI said last. - The teacher remains the orchestrator.
In Rotella’s room, the teacher nudges, redirects, invites, and sometimes gets out of the way. Topoi is designed so that teachers define roles, set expectations, and shape how AI is allowed to participate. The AI doesn’t grade, decide, or set the agenda.
In other words: if Rotella’s class is a band rehearsal, Topoi’s job is to be the sound system, not a replacement musician.
Where Our Approaches Meet
Rotella’s essay is full of moments that could map directly into a Topoi-style environment.
- When students co-design exam questions at the end of class, that’s a perfect moment for a shared Topoi session: “Let’s ask the system to propose three possible exam prompts based on today’s discussion—and then critique them together.” The point isn’t to accept its suggestions, but to sharpen students’ sense of what a good question even is.
- When a teacher wants to emphasize close reading over summary, AI can be positioned as a challenger rather than a shortcut: “Here’s the bot’s summary of the story. Where is it missing texture, irony, or ambiguity that you noticed?” The comparison can deepen attention, not replace it.
- When students are working through drafts, the scaffolding Rotella describes can be documented and supported in a shared space, with AI helping to surface patterns (“Here are three claims you’ve been circling around”) that the class then evaluates.
None of this requires giving up the blue book, the pen, or the device-free room. It simply means that when AI is used, it’s used in public, under teacher guidance, and in the service of the same “equipment for living” Rotella cares about.
Beyond Resistance: Fluency Without Surrender
Near the end of the essay, Rotella compares school to a gym: machines can do the lifting, but if you want to get stronger, you still have to move the weight yourself. Students, he argues, need teachers who show them how to use machines properly—and teachers who show them how to do the work unassisted.
Topoi is very comfortable in that analogy.
It doesn’t want to be the machine that lifts for you. It wants to be the mirror, the rack, the whiteboard with your workout written on it, and maybe the coach who occasionally points out that your form is off. The effort is still yours. The “burn” still belongs to the student. The difference is that the environment has been intentionally designed to make that effort visible, shared, and meaningful.
Rotella’s “AI-resistant” classroom is, in a sense, already doing Topoi’s work without the software:
- Centering process.
- Honoring attention.
- Building community.
- Protecting the slow, effortful parts of thinking.
Topoi doesn’t ask him—or teachers like him—to stop doing any of that. It asks:
What if the tools we use for AI were redesigned so that they reinforced those choices instead of undermining them?
That’s the experiment this project is trying to run.
The Real Question
Tyler’s question still rings: How do we learn to ask the questions on our own?
Rotella’s answer is to keep building classrooms where students practice exactly that, together, with books and pens and voices in the same room. Topoi’s answer is compatible:
- Preserve that room.
- Strengthen that community.
- And when we bring AI into the picture, make sure it’s there to help humans ask better questions—not to answer them in our place.
That’s not the AI apocalypse. It’s a chance to rebuild the tools of learning so they actually serve what teachers like Rotella have been doing all along.