Learning Is a Struggle AI Must Not Skip (NotebookLM Version)

Welcome to the Cultural Context of Knowledge podcast. I'm your host, Donald Easton-Brooks. This podcast invites learners to pause and reflect on how their own cultural experience shaped the way they learn, interpret information, and define what counts as knowledge. Each episode will focus on how cultural context impact the way in which we understand the world around us, whether it's the classroom work environment, our relationships.

In this episode, learning is a struggle. AI must not skip. I will be using NotebookLM. This is an AI program that allows you to put in your notes and it will create a podcast using AI voices. This program is very good. I had taking notes and conceptualizing it into a two person conversation. I would do a more in depth conversation on NotebookLM, and a later episode.

This episode will focus on learning and the scaffolding of learning and how learning is impacted by ai. If it's not used effectively, the program will also focus on how culture plays an important role in knowledge and how knowledge and culture is important as we continue to explore ai. And now present to you notebook, lms interpretation of my work, centered on the culture context of knowledge, and the impact of AI on learning.

Welcome to the deep dive. This session is really custom tailored for you, for our audience of educators, teachers, professors, all of us who are trying to figure out how to handle these incredibly powerful AI tools. You know, like chat GPT in the classroom. Exactly. And today we're diving deep into the conceptual framework.

You really need for that. Drawing a lot from the work of Dr. Donald Easton-Brooks. The core mission here is, uh, to square this circle. Really. How do we reconcile the fundamental nature of learning? The fact that it's a necessary struggle with tools that seem to offer an easy way out. That's the conflict right there, isn't it?

AI gives you this promise of instant answers and efficiency, but if we just, if we just grab these tools without thinking about the pedagogy of the foundation, we risk undermining the whole process, right? We could be sabotaging the very way deep knowledge is built. It's like we're trying to use a new power tool without destroying the actual craft of woodworking, and the sources are very clear on this.

The danger isn't the AI itself, it's how we use it. Learning is this, uh, profound sequential thing built on cognitive friction. If we let students bypass that essential struggle, and what do we get? You get students who can generate output, but they don't have the underlying understanding to actually critique it, to evaluate it, or to apply it in any meaningful way.

Okay, so let's unpack that, that core idea learning is struggle. The article has a beautiful definition for this. It's the process of of working through information that was, uh, ambiguous or unknown until your brain finally gets to a state of equilibrium that wrestling with a concept. That's the engine and the article points out there are different kinds of struggle, right?

It's not all the same. Right? And it's differentiated based on where the learner's coming from. The article details two main types. Okay? First up is the simple struggle. Yeah. This is when a task is completely doable, but you know, the learner just hasn't done it before. They don't have the experience. The article uses this great example of a kindergartner learning to write the letter a.

Right. They can hold the pencil. They have the motor skills. Exactly. But forming those two specific diagonal lines in that little crossbar that takes rehearsal, right? It takes repetition until it becomes automatic. The struggle there is purely experiential. It's the novelty, the coordination of it all.

Okay, so how is that different from the complex struggle? Well, the complex struggle requires you to already have some knowledge. A base level of skill has to be there before you can even start. So like trying to learn calculus if you never really got algebra. Perfect example. Yeah. Or trying to write a complex essay before you've mastered how to form a basic sentence.

The level of struggle is totally based on the learner's frame of reference. If that foundation isn't there, it's not productive struggle, it's just. Frustration in this whole idea of building blocks of a sequence. It leads us right to Bloom's taxonomy. The article frames it as scaffolding, right? Not a rigid ladder, but you have to build things in order.

You absolutely do. You have to build the blocks in a certain sequence. So let's break down the three fundamental stages from Bloom's that the article really focuses on. The first one is. The basic level, the foundation, non-negotiable. Yeah. This is where you develop that fundamental understanding. You know, definitions, basic functions, the what and the why of the knowledge.

And for you as an educator, this is so critical because motivation just plummets. If a student doesn't see the point of this basic stuff, you have to get that baseline down before you can move on. Okay, so what's next? Next up is the compare and contrast level. So connecting the dots. Exactly. Understanding gets deeper here because you're relating the new knowledge to things you already know and you're distinguishing between them.

It's not just knowing what the letter A is, it's comparing its look and sound to B or in science, not just memorizing photosynthesis, but comparing it to cellular respiration. Building a map, you're building that conceptual map, and once that map is built, then okay, only then you get to the top, to the abstract level, and this is where the magic happens.

This is where the creativity, the complexity, the higher level thinking lives, but it completely depends on a solid mastery of those first two levels. You can't build a skyscraper on a foundation of sand. Okay? Not this. This is where it gets really interesting for every educator thinking about ai. The article makes this powerful point that when people use tools like Chat, GPT, they almost always start at the third level, the abstract, highest level.

They skip the struggle. They just ask for the finished product. Write me an essay, gimme a policy proposal right outta the gate. And the article says that is the number one conceptual flaw in how we're integrating ai. Okay, but hang on, let me push back on that a little. The article says backward learning is possible.

So what if I'm a student and I ask Chat GPT for that final complex answer, say, analyze the socioeconomic impacts of the 1990 Clean Air Act. And then I spend the next three days going backward, researching every term, every citation. That AI response isn't that, isn't that a valid kind of struggle? That's a great point.

And yes, that can happen. But the article's position is that it's just dramatically less efficient. The problem is you're deconstructing a final product without the framework to even know it's important. You're trying to learn level one and two concepts by dissecting a level three answer, it's a much longer, more painful process.

The AI is just a predictive model. It's designed to give the most statistically plausible answer, not necessarily the most conceptually accurate one. For my specific learning goal, precisely for that AI to be a real partner, a copilot, you have to provide the direction. You need that conceptual knowledge.

If you don't have it, the AI just spits out something generic and plausible, and you waste a ton of time. Let's get really practical. Let's detail the pitfalls that happen when a student gives a vague prompt because they don't have that conceptual direction. These are really critical warnings for all of us to pass on.

The first big one is just an inefficient process. Yeah. It turns AI into a huge time sink. What does that look like in a real assignment? Well, imagine a student needs to write a report on supply chain. If they don't know the basic difference between say, procurement and logistics, that's a level one concept.

Their first prompt might be something vague, like right about getting stuff from China to the us. Okay? The AI gives them a general report, but they see it's too broad. So now they spend hours just tinkering with the prompt, make it more focused, add some data. They're doing all this trial and error, instead of just spending 10 minutes learning the core definitions they missed.

That's just unproductive struggle. Right. And the second pitfall is getting irrelevant or overly general output. Yeah. This is when the AI gives you high quality information, but it's completely wrong for your specific need. A history student asks for a paper on the causes of World War I, but they haven't grasped the importance of the alliance system.

The AI might give them a brilliant paper on industrial competition, which is true, but it's not what the assignment was about exactly. It completely misses the point. So for the task at hand, it's basically useless knowledge, which leads right to the third issue. Difficulty in evaluation. How can you possibly know if the AI's response is any good if you don't have that?

Level one and two, understanding, you can't tell a brilliant answer from a flawed one. So you might end up submitting something that just sounds authoritative, but is actually really low quality. Okay, and the fourth, this is the one everyone talks about, the increased risk of errors. The AI hallucinations, a specific conceptually driven prompt acts as a guardrail.

It keeps the AI in a defined space. But when the prompts are vague, tell me about history. The guardrails are gone and AI has more leeway to just. Make stuff up. Plausible sounding stuff, fabricated information, fake sources, bad data, huh? And the student, because they don't have the conceptual knowledge, has no way of knowing.

It's a complete fabrication. And the last one here is the output results in a lack of actionable next steps. The knowledge just sits there. It's inert. If the student didn't know what they were gonna do with the answer in the first place, the output is just text on a screen. True learning means taking that knowledge and applying it.

Using it to solve a problem. So this takes us from the technical problems of AI to something much deeper, much more philosophical, even if we could perfect the prompts. The article raises this massive challenge. AI has not perfected the cultural context of knowledge. We have to look beyond just the facts.

And this is a fascinating shift in the article. It moves from the hard sciences, you know, with absolute laws like gravity to the social sciences and in the social sciences, things are governed by the law of human response. Human experience as the article quotes has context. There is always a history, and who tells that history, whose context matters.

This gets us to the crucial difference between dominant and non-dominant cultural knowledge. Something every educator has to grapple with. We have to talk about grand narratives versus many narratives, grand narratives. They're the stories generated by the dominant culture, usually the one with wealth and control.

And this creates a biased framework that shapes everything. From policy to curriculum, it creates these neat little boxes, right versus wrong, good versus bad. And while it's building those neat boxes, it pushes out other stories. It rejects the many narratives, the beliefs, the lived experiences of minority groups, of underclass groups.

And the article argues this rejection. Leads to huge systematic flaws, even what it calls dehumanization. Can you give us an example of how that works? Sure. Think about a standard high school history textbook. The grand narrative might be about the economic boom of the 1950s, GDP growth, new houses in the suburbs.

It's a story of order in progress. Okay. But the many narratives, the stories of marginalized communities who are systematically cut outta that boom through things like redlining, those stories are often just. A fur note or they're not there at all. So the textbook, by only showing the dominant view, is giving you incomplete knowledge, which is why we as educators have to push beyond that.

The article brings up two really interesting epistemological tools to fight this pragmatism and fallible. Pragmatism is basically about accepting that there's no single, perfect model that explains everything. You can have multiple, even contradictory models that all help explain one phenomenon. So for an educator, that means teaching students that if you only look at something through one lens, one economic theory, one historical perspective, your answer's probably wrong.

Or at least incomplete. You have to look at all the influences. And what about fallible? Fallible is even more radical. It challenges the whole idea of certainty. It says that all knowledge we currently have, even stuff we think of as hard facts, could be an error. So we should just throw everything out. No, no.

That would be skepticism. Fallible says we should accept prior knowledge is useful, but it has to be open to revision based on new evidence. The classic example is the shift from Newton's laws of gravity to Einstein's theory of relativity. Newton wasn't wrong. He was just incomplete. Fallible is asks us what are we incomplete about today and bringing this all back to the classroom.

The article really emphasizes the educator's role in building this collective knowledge. It requires constant negotiation looking at a situation from every single angle. Absolutely. And if we exclude those mini narratives. Our knowledge is incomplete and our solutions will fail. The article's, example of the Civil Rights Act and school busing is just absolutely so stark.

Mm-hmm. The policy makers, the dominant culture, they failed to account for the many narratives of the parents, the teachers, the actual communities involved. So even with good intentions, because the policy was based on incomplete knowledge, it had all these implementation flaws and negative consequences.

You have to collect all the angles. Yeah, all the narratives. So the message, the pedagogical imperative for everyone listening is pretty clear. We have to adopt a multi-dimensional framework. We have to teach our students not just how to use ai, but how to think conceptually, how to climb Bloom's hierarchy so they can be the ones guiding the tool.

We have to train them to be skeptical, but also pragmatic, to take apart the grand narratives, to insist on finding the mini narratives and to always be ready to accept that what they know now might be incomplete. That's the only way they can turn AI from a shortcut into a truly powerful analytical tool.

That's the synthesis right there. The output from an AI is only as good as the conceptual and cultural framework that the user brings to it. If your basic understanding is weak, the AI is useless. If your cultural understanding only includes the grand narratives, the AI will just reinforce biases and flaws.

Which brings us to our final provocative thought drawing on pH Collins work from the article. Collins suggests that for oppressed communities to gain real power, they must actively debate the theories related to them and shift the whole paradigm away from the dominant culture's lens. Now AI is this incredibly powerful, culturally neutral engine.

So if we see it as a tool that can synthesize huge amounts of scattered data, all those mini narratives out there, how can you as an educator, harness this tool to facilitate that exact debate in your own classroom? How can you use it to ensure those suppressed voices aren't just heard, but are actively used to build a truer, more complete collective knowledge?

This transcript was generated from the episode recording and may contain small transcription errors.