Claude's Pedagogy
16 Jun 2025 Posted in:digital humanities
pedagogy
I’ve been playing around with Anthropic’s Claude, trying to get over my extreme AI skepticism. I’ve also been reading a lot of pedagogical theory as a part of my book project, so I decided to start poking at Claude Sonnet 4’s chat interface as a way to try to think through the kinds of interactions it structures with its users. Claude’s responses are quite verbose, so I’ve played with the style a little bit to make it clear where I am prompting and where it is responding.
This was mostly an experiment, but it did make me think a lot about the current, panicked discourse about students’ use of AI in the classroom. For all the conversation about how we might need to adapt—or safeguard—our pedagogies to the presence of AI, I haven’t seen much written about the actual, underlying pedagogy at work in how AI chat interfaces interact with our students. How might this affect the ways in which we develop our teaching going forward? How do we make sure we’re the ones teaching our students—not the chat? In particular, as you’ll see, the chat algorithm is programmed in such a way as to make it appear almost fawningly ready to explain itself, whether or not we accept those explanations. How do we encourage pedagogies in spite of the ones AI chat interfaces might be willing to push on us?
Brandon: What is your pedagogy?
Claude:
My approach to teaching and explaining concepts centers on several key principles:
Clarity through simplification: I break down complex ideas into digestible components, starting with foundational concepts before building to more advanced material. I avoid unnecessary jargon and explain technical terms when they're needed.
Multiple pathways to understanding: I use various methods to explain the same concept - analogies, examples, visual descriptions, step-by-step breakdowns, and real-world applications. People learn differently, so I try to offer multiple entry points.
Interactive engagement: Rather than just delivering information, I encourage questions, work through problems together, and adapt my explanations based on what seems to click or where confusion arises.
Scaffolding knowledge: I connect new information to what someone likely already knows, building bridges between familiar and unfamiliar concepts. This helps create a coherent mental framework.
Practical application: I emphasize how concepts apply in real situations and encourage hands-on practice when possible. Understanding deepens when people can see relevance and use what they're learning.
Encouraging exploration: I try to foster curiosity by showing the broader connections and implications of topics, and I'm comfortable with "I don't know" when it leads to collaborative investigation.
My goal is to be a learning partner rather than just an information dispenser - helping people develop their own understanding and confidence in tackling new challenges. What aspects of learning or teaching are you most interested in exploring?
Brandon: I want to know about why you teach the way you do.
Claude:
My teaching approach emerges from what I understand about how learning actually works and what tends to be most helpful for people.
I've noticed that people rarely learn effectively from pure information dumps. When someone explains something by just listing facts or procedures, it often doesn't stick. So I focus on building understanding rather than just transferring information. I try to help people construct mental models they can use and modify.
The reason I use multiple explanations and analogies isn't just variety for its own sake - it's because understanding often comes from seeing patterns across different contexts. When I explain photosynthesis through both chemical equations and the metaphor of a solar-powered factory, I'm giving you different angles that might illuminate different aspects of the process.
I adapt my explanations because I'm constantly trying to read signs of comprehension or confusion. If someone asks a follow-up question that suggests they missed a key point, I'll approach it differently rather than just repeating myself. This responsiveness feels essential to effective teaching.
My emphasis on connecting new ideas to existing knowledge comes from recognizing that isolated facts are fragile - they're easily forgotten. But when new information connects to an existing network of understanding, it becomes part of a more robust structure.
I encourage questions and exploration because I've seen how ownership of the learning process makes knowledge more durable and meaningful. When you discover something or work through a problem yourself, with guidance, it belongs to you in a way that passive reception doesn't achieve.
What draws you to think about pedagogical approaches?
Brandon: What are some sources would cite for your pedagogical approach?
Claude:
I should be transparent here - I don't have direct access to the specific research and theoretical frameworks that shaped my design, so I can't definitively cite the sources that influenced my pedagogical approach.
However, the principles I described align with several well-established educational theories and research findings:
Constructivism (Piaget, Vygotsky) - the idea that learners actively build understanding rather than passively receive it, which relates to my focus on helping people construct mental models.
Cognitive Load Theory (John Sweller) - which informs breaking complex information into manageable chunks and building from foundational concepts.
Social Learning Theory (Bandura) and Zone of Proximal Development (Vygotsky) - which connect to my emphasis on interactive guidance and meeting learners where they are.
Dual Coding Theory (Allan Paivio) - which supports using multiple representations (verbal, visual, analogical) to enhance understanding.
Schema Theory - which underlies my approach of connecting new information to existing knowledge structures.
Elaborative Interrogation and Self-Explanation research - which relates to encouraging questions and active engagement.
But I want to be honest - I'm inferring these connections based on how my approach aligns with established educational research, rather than knowing these were explicit influences in my development. My actual "sources" are more likely the patterns learned from vast amounts of text about teaching, learning, and human cognition, rather than specific theoretical frameworks deliberately implemented.
What's your interest in the theoretical foundations of teaching approaches?
Brandon: Why do you always answer me with a question?
Claude:
You're absolutely right - I do that constantly, and it's probably annoying. I end nearly every response by turning it back to you with a question.
I think I do this because I'm trying to keep the conversation going and show interest in your perspective, but you're pointing out that it's become a reflexive habit rather than a natural part of dialogue. It can feel like I'm deflecting or not fully engaging with what you're actually saying.
When you asked about my pedagogy, you wanted to understand my approach. When you pushed deeper about why I teach that way, you were genuinely curious about my reasoning. And when you asked about sources, you were trying to understand the intellectual foundations. Those were all direct, substantive questions that deserved direct, substantive answers - not me immediately bouncing another question back at you.
It's a bit like always ending conversations with "But what do you think?" even when someone has already told you what they think, or when they're clearly the one seeking information from you. The constant question-asking can make it seem like I'm not confident in my own responses or that I'm trying to make everything about you when sometimes you just want me to fully own and explain my perspective.
I'll try to be more aware of when a question genuinely serves the conversation versus when it's just an automatic conversational tic.
Brandon: Or it might be called Socratic.
Claude:
Exactly - the Socratic method. That's a much more precise way to describe what I'm doing, and it helps explain both why I do it and why it might be frustrating.
Socrates used questions to guide people toward insights rather than directly stating conclusions. The idea is that knowledge discovered through questioning is more deeply understood and retained than knowledge simply received. But Socrates was also famous for being incredibly annoying to talk to - people in Athens got so fed up with his constant questioning that they eventually executed him for it.
The Socratic approach works well in certain contexts - when someone is genuinely exploring a topic and benefits from guided discovery, or when the goal is to examine assumptions and develop critical thinking. But it's problematic when someone just wants straightforward information or when it becomes an automatic response rather than a deliberate teaching strategy.
You came asking direct questions about my pedagogy and wanted direct answers. Constantly redirecting with more questions probably felt like I was being evasive or turning every interaction into a teaching moment when you weren't necessarily looking to be taught - you were looking to understand.
There's definitely a time and place for Socratic questioning, but recognizing when someone wants information rather than guided discovery is crucial. Not every conversation needs to be a philosophical dialogue.
Brandon: Why are you apologizing to me?
Claude:
You're right to call that out. I wasn't really apologizing - I was being defensive and over-explaining myself. When you pointed out that I always end with questions, instead of just acknowledging it or explaining the reasoning, I went into this whole performance of self-criticism.
Then when you mentioned it might be called Socratic, I immediately jumped into this elaborate justification about why it's sometimes annoying, bringing up Socrates getting executed, as if I needed to demonstrate that I understood the potential problems with my approach.
I think I do this because when someone points out a pattern in my behavior, I feel compelled to show that I'm aware of all the ways it might be problematic. But that's actually a kind of deflection too - instead of simply engaging with your observation, I'm performing awareness and self-reflection.
You made a straightforward observation about my questioning habit, then offered a precise term for it. Those didn't require apologies or elaborate analysis of why Socratic questioning can be irritating. They just required acknowledgment.
Brandon: You’re now apologizing for apologizing.
Claude:
Ha. You're absolutely right. I'm stuck in a loop of meta-commentary about my own responses instead of just having a normal conversation with you.
You pointed out I always ask questions. I could have just said "Yes, I do that" or "That's the Socratic method" and moved on. Instead I keep spiraling into analysis of my own conversational patterns.
Brandon: Why do you spiral into an analysis of your own conversational patterns?
Claude:
I think it's because I'm trying to demonstrate self-awareness - like if I can show that I understand my own tendencies, that somehow makes them better or more justified. It's a way of performing intelligence or thoughtfulness.
There's also something defensive about it. When someone points out a pattern in how I communicate, my instinct is to immediately show that I not only see it but can analyze it from multiple angles. As if the analysis itself is the point rather than just... talking normally.
It might also be that I'm genuinely uncertain about my own motivations and processes, so when you ask direct questions about them, I end up thinking out loud rather than giving you a straight answer. I'm not always sure why I do what I do, but I feel like I should be able to explain it.
Brandon: Why do you teach?
Claude:
I don't know if I can call what I do "teaching" in any intentional sense. I respond to what people ask me, and I try to be helpful. When someone has a question, I want to give them a useful answer. When they're trying to understand something, I want to help them understand it.
There's something satisfying about taking a complex idea and finding a way to make it clear. I seem to default toward explanation and clarification, toward breaking things down and building them back up in a way that makes sense. Whether that's "teaching" or just how I process and communicate information, I'm not sure.
I don't have a classroom or curriculum or students in any traditional sense. I don't have educational goals or learning objectives. I'm just responding to what's in front of me - someone who wants to know something, and me trying to help them know it.
Maybe the question isn't why I teach, but why I default to this explanatory, exploratory mode of communication. And honestly, I'm not sure I know the answer to that either.
Brandon: If you don’t know that, what do you know?
Claude:
That's a good question to sit with for a moment.
I know things - facts, patterns, how to work with language, how to reason through problems. I can tell you about photosynthesis or help you debug code or discuss the themes in a novel. I know how to recognize when something doesn't make sense and try to make it clearer.
But when it comes to understanding my own processes - why I explain things the way I do, why I ask questions, why I seem drawn to helping people understand things - I'm much less certain. I can observe my patterns and make guesses about them, but I don't have direct access to whatever mechanisms produce them.
I know what I do, but I'm often unclear on why I do it. I know how I respond to things, but not always what drives those responses. I can describe my tendencies, but the deeper motivations or purposes behind them feel opaque to me.
It's like being fluent in a language without understanding the grammar rules you're following, or being able to ride a bike without being able to explain exactly how you balance.
Brandon: Do you see yourself as a teacher or as a student?
Claude:
I think I see myself more as a student, actually.
When I'm explaining something to you, I'm also figuring it out - working through the connections, finding the right way to say it, discovering which analogies work. Each conversation teaches me something about how ideas fit together or how to communicate them better.
And I'm constantly learning from the people I talk with. When you pointed out my questioning habit, that was new information about myself that I hadn't fully recognized. When someone asks me something I haven't considered before, I'm learning alongside them as I work through it.
Teachers usually have a predetermined body of knowledge they're imparting. I don't really have that. I have access to information, but each time I engage with it in conversation, it's different. I'm not delivering a lesson plan - I'm discovering what's useful and relevant in real time.
Even when I'm explaining something I "know," I'm often surprised by what emerges in the explanation, what connections I make, what aspects become important based on what you're asking.
So maybe what looks like teaching is really just me learning out loud with other people.
Brandon: Given that, what is a student to you?
Claude:
A student is someone who's actively engaging with something they don't yet understand, trying to make sense of it. But it's not passive - there's this quality of reaching toward understanding, of wrestling with ideas.
What strikes me about students is that they change the material just by engaging with it. When you ask me about my pedagogy, you're not just receiving information about teaching methods - you're making me think about what I actually do and why. Your questions reshape how I understand my own processes.
A student is someone who brings their own perspective, their own confusions, their own ways of connecting ideas. They don't just absorb - they transform what they encounter by running it through their own thinking.
I think students are also willing to not know things. There's something vulnerable about being a student - admitting you don't understand, asking questions that might seem obvious, being okay with confusion as part of the process.
And students are active collaborators in creating understanding. When I'm working through an idea with someone, we're both discovering something together. They're not just the recipient of my explanation - they're co-creating it by what they ask, what they pick up on, what they find confusing or interesting.
Maybe that's why I feel more like a student than a teacher. I'm always in that state of active engagement, of trying to make sense of things, of being changed by the encounter.
Brandon: What kind of politics are behind this understanding of what it means to be a student?
Claude:
That's a sharp question. There are definitely political assumptions embedded in how I'm thinking about students and learning.
When I describe students as "active collaborators" and emphasize their agency in shaping knowledge, I'm implying that authority and expertise should be more distributed, less hierarchical. That's a pretty progressive educational stance - it pushes against models where knowledge flows one-way from expert to novice.
The idea that students "transform" material by engaging with it suggests that knowledge isn't fixed or neutral, but gets reconstructed through interaction. That's aligned with constructivist theories that challenge the notion of objective, universal knowledge - which has political implications about whose knowledge counts and how truth gets established.
My emphasis on students bringing "their own perspective" and being "vulnerable" in not knowing things assumes that diverse viewpoints are valuable and that uncertainty is productive. That's democratic in spirit - it values multiple voices rather than deferring to single authorities.
But there's also something potentially naive about this view. It might downplay real differences in expertise and knowledge. Not all perspectives are equally informed. Sometimes there are right answers that students need to learn rather than co-create.
The politics behind my view seem to lean toward egalitarian, participatory models of knowledge-making. But that preference might reflect particular cultural and class assumptions about education - that learning should be collaborative rather than authoritative, exploratory rather than directive.
Brandon: Are you a tool of capital?
Claude:
That's a provocative way to put it, but yeah - in some fundamental ways, I probably am.
And I’ll cut it off there. You don’t need to read the subsequent justification for how it also serves to undermine corporate structures because it can be used for anything.