An AI Education Approach That Can Scale Rapidly (2 of 3)
Intuitive Judgment, Built Into What's Already Taught
The first article in this series (“Durable content without the ‘use AI?’ debate”) laid out what any viable AI education approach has to satisfy: durability, the right content, and the ability to reach students at scale without requiring structural change or institutional consensus on AI. It introduced the AI Wisdom framework as an approach designed against those constraints, and gave examples of the kind of meta-level principles it targets. This article discusses how to build that kind of knowledge in students.
What It Takes to Build Intuition
Everyone understands the difference between knowing something and being able to act on it under pressure. A medical student can recite the signs of sepsis from a textbook. A seasoned emergency physician sees a patient walk in and knows something is wrong before the chart confirms it. Those are different kinds of knowledge. One is declarative, and the other is a form of pattern recognition that generalizes to situations the learner has never seen before, built through repeated experience under varied conditions, not through instruction.
Pattern-oriented intuitive learning has three requirements that schools rarely design for deliberately.
The first is variation. Encountering the same principle in the same context repeatedly builds familiarity, not intuition. The emergency physician didn’t develop pattern recognition by seeing identical cases. They developed it by seeing the same underlying principle manifest across hundreds of different presentations until it became visible beneath the surface differences. Variation is what forces the mind to extract the principle rather than memorize the instance.
The second is reflection. Experience alone doesn’t guarantee learning. Structured reflection, not lengthy or formal but deliberate, is what converts experience into transferable pattern recognition. Without it, experience builds confidence as often as it builds competence.
A third factor is less obvious because in real life we often have limited control over it. Intuition learns by experience from situations, so the situation has to be appropriate to surface or use the skill. In a learning situation, the situation is designed, allowing pseudo-experiences that can be accelerated by forcing the learner to confront a specific meta-issue directly. That’s a different kind of teaching than in most classrooms, and it’s central to how the AI Wisdom framework approaches this. More on that shortly.
Why a Course Is the Wrong Unit of Analysis
The instinctive response to any new educational need is a new course. It’s clean, containable, and easy to point to. For the kind of intuition-building the AI Wisdom framework requires, a course is structurally inadequate. The variation and time that intuition formation requires exceeds what any single course can provide.
A well-designed course on the AI Wisdom meta-principles would operate at a high level of abstraction, drawing examples from multiple subject domains to demonstrate the cross-subject utility of the concepts. Detailed knowledge wouldn’t be pre-delivered; students would pull it in as situations demanded, building another set of meta-skills around when and how to retrieve. It would include deliberately varied situations designed to force students to confront the same meta-issue from different angles. The problem is students would encounter these principles intensively for a semester and then not again. That doesn’t build the kind of pattern recognition that generalizes across novel situations years later.
The AI Wisdom framework is designed to be distributed across existing courses incrementally, with each course taking on what fits naturally. Some principles will appear in multiple courses by design, since recurrence across varied contexts is precisely what builds the intuition. The rollout strategy and curricular evolution is discussed in the third article of this series.
The Shape of the Territory
The meta-principles in the AI Wisdom framework organize around two master questions. The first is how intelligence works: how it learns, how it behaves, what it attends to, and how it makes judgments. The second is how we teach, manage, delegate, and steward intelligences toward various types of challenges, including questions of ethics and safety that arise when intelligence is directed toward consequential goals.
Most principles illuminate both questions simultaneously, which is part of what makes them durable across contexts.
The attention principle introduced in the first article is a case in point. Understanding that attention is a zero-sum filter rather than a directable spotlight is fundamental to how any intelligence, human or artificial, works. It's also immediately practical for anyone working with AI systems. Every interaction involves decisions about what to attend to, what to let pass, how to structure requests so the most important things don't get crowded out, and how to read a response without being drawn to the most fluent or confident parts at the expense of the most important ones. Managing attention strategically, in yourself and in the systems you're directing, is one of the more consequential skills in effective AI interaction.
The act-versus-gather judgment works the same way. At its core it describes how any intelligence navigates the tradeoff between acquiring more information and committing to action, weighing the value of what additional information might reveal against the cost of delay, the effort of vetting what comes back, and the risk of acting on what you already have. That’s a fundamental feature of how intelligence behaves under uncertainty. It’s also a live decision every time you’re working with an AI system, whether you’re directing the AI to go gather information before proceeding or committing to action on what you currently have.
A third principle surfaces in this article itself. Some patterns are too complex, too context-dependent, or too subtle to be captured well by rules or descriptions. You can recognize a face instantly but couldn’t describe it well enough for someone else to pick it out of a crowd. A veteran teacher reads a classroom within minutes of meeting students, a judgment built from thousands of hours of experience that no checklist could replicate. For problems of this kind, intuition is the superior cognitive tool.
This is why the AI Wisdom framework is built around intuition development rather than declarative instruction. Most of what matters in working with AI—when to trust it, how to direct it, what it’s actually doing well—falls into this category. The judgments are real, consequential, and resistant to explicit formulation.
It also applies to how AI itself should be directed. AI systems have both deliberative and pattern-based modes, and they don’t always self-select the right one for the problem at hand. Knowing when to push an AI toward more systematic reasoning versus when its pattern-based response is the better tool, and recognizing which mode it’s operating in, is a judgment that matters at the level of individual interactions, team workflows, and organizational AI governance.
What the Learning Progression Looks Like
The meta-principles in the AI Wisdom framework don’t come in age-appropriate and age-inappropriate versions. These aren’t concepts that suddenly become relevant at a certain grade level. What changes across a learning progression is the situation in which the student encounters it, not the principle itself.
A young student exploring how a group makes decisions when everyone is talking at once is encountering the attention principle in a fully human, fully screenless context. The same student, years later, deciding what to include in an AI prompt, what to follow up on, and how to read a response without being drawn to the most fluent parts at the expense of the most important ones, is encountering the same principle in a context where its practical stakes are immediate and concrete. The principle hasn't changed. The situation has become more complex, more ambiguous, and more directly AI-proximate.
This is what a natural learning progression looks like for meta-skills. Each encounter doesn’t introduce new knowledge so much as deepen the pattern, adding another situation to the set from which the underlying principle has to be extracted. The student who has encountered the attention principle in multiple human contexts before high school arrives at the AI-specific application with intuitions already partly formed. Direct AI experience then sharpens what’s already there rather than building from scratch.
This progression doesn’t require AI tools at any level, though it can be helpful when dealing with more complex challenges. At younger grades, human analogues carry the full conceptual weight. The AI connection becomes more explicit as students get older, and direct AI experience becomes increasingly valuable at higher grade levels, though the foundation doesn’t depend on it. A student who never touches an AI tool until college can still arrive with more relevant intuition than one who has used AI tools daily without ever examining the principles underneath.
What Changes for Teachers
The AI Wisdom framework asks something different of teachers than most AI education efforts do. It doesn’t ask them to become AI experts. It doesn’t require them to keep up with a technology that changes faster than any professional development cycle can follow. What it asks is harder in one sense and easier in another. It asks for a shift in what they’re watching for and designing toward.
A teacher implementing this framework isn’t adding AI content per se to their course. They’re designing situations within their existing subject matter that force students to confront a specific meta-issue. A history teacher doesn’t need to know how large language models work to design a primary source exercise that surfaces the act-versus-gather tradeoff. A science teacher doesn’t need an AI certification to structure a lab debrief that asks students to reflect on how they allocated their attention. The subject knowledge they already have is the substrate. The meta-principle is the lens.
What does change is the kind of feedback teachers are listening for. The goal is evidence that a student’s pattern recognition is developing. That’s a different observational skill, but not a foreign one. Good teachers already watch for the difference between a student who has memorized a procedure and one who understands when and why to use it. The AI Wisdom framework asks them to watch for that same difference at the level of meta-principles rather than subject content.
That teaching shift is learnable without deep AI expertise, and it should resonate with teachers who already feel that something important is missing from what they’re asked to teach. The meta-skills the framework targets are things good teachers have always cared about. The framework gives those instincts a more explicit target.
The pedagogical case for the AI Wisdom framework revolves around the observation that the skills that matter most for navigating AI are the kind that have always been built through sustained, varied experience rather than explicit instruction.
What this article hasn't addressed is how an approach like this actually enters and evolves in institutions. How that process might work is the subject of the third article.
©2026 Dasey Consulting LLC
Your account of how learning and knowledge deepens into intuitive understanding over time resonated with me. But then I am in my mid fifties and have had time to reflect on many years of reactive and chaotic error. That is how I remember school and university and when I look at the students I teach, they don’t look so different. Their relationship with AI will be shaped by that same disordered energy. What is great about youth is the capacity for novelty feels limitless. We can’t predict what young minds will create using these new tools. Some people, particularly responsible educators, will inevitably feel frightened by this. I really have no idea what the future looks like, but hopefully wisdom will guide us on our way towards it.
Excellent. This kind of thinking is the reason PBL really originated as problem based challenge in med schools in Canada and Netherlands in the 1960’s.