Why Expertise Disappears When Institutions Stop Training the Next Generation

Jul 18, 2026

10 min read

2 views

87%

The Hidden Fragility of Human Competence

What happens when a culture stops passing on formation, but still expects the next generation to perform as if it had been formed?

That question sounds religious on the surface, but it reaches far beyond religion. It helps explain why some families lose faith across generations, why organizations fill with people who can use tools but not truly master them, and why so many modern systems mistake exposure for apprenticeship. We are living in a world that assumes continuity while quietly dismantling the conditions that make continuity possible.

The result is a strange kind of institutional amnesia. Parents assume children will absorb what they once absorbed. Managers assume AI or onboarding will make up for thin expertise. Schools assume information access is the same thing as education. But in each case, the deeper reality is the same: competence does not reproduce itself automatically. It has to be intentionally transmitted, protected, and embodied.

The most dangerous assumption in modern life is that if we preserve the appearance of a practice, we have preserved the practice itself.

This is not just a story about belief or technology. It is a story about how humans become the kind of people who can sustain a tradition, a craft, or a responsibility. And once you see that pattern, it becomes hard to unsee it.

The Illusion of Inheritance

Many parents, leaders, and institutions live by an invisible script: do roughly what the previous generation did, and the next generation will turn out roughly the same. That script used to have some truth to it because the surrounding environment reinforced it. Children grew up in a world where the family, the school, the church, and the neighborhood all pointed in similar directions. Formation was not perfect, but it was dense.

Now that density has thinned. Children spend more time in fragmented media environments, less time in shared communal spaces, and more of their lives in institutions that do not assume any particular moral, spiritual, or practical inheritance. They may still hear the language of the past, but they are no longer immersed in its habits. In other words, many adults are trying to pass on a way of life without passing on the ecosystem that made that way of life feel inevitable.

This explains a great deal about religious decline, but it also explains something broader. The modern world increasingly treats identity as a matter of preference and skill as a matter of access. If you can stream the content, search the answer, or ask the model, why would you need deep initiation? Yet initiation is exactly what gets lost when a system relies on shortcuts.

A child can hear an answer to a theological question, just as a junior employee can ask an AI for a marketing plan. But hearing an answer is not the same as being formed by a discipline. The child may know what words to use. The employee may know what output to request. Neither necessarily understands the moral weight, the interpretive frame, or the hidden structure underneath.

The core mistake is confusing availability with internalization.

AI and the Apprenticeship Gap

Generative AI exposes this problem in a new way. It is often remarkable at accelerating work, suggesting options, and shrinking learning curves. It can help a data scientist move into another role, outline ideas for a new project, or assist someone in performing a task they do not do every day. In that sense, it functions like a force multiplier for people who already possess some domain knowledge.

But there is a hard ceiling. When a person lacks sufficient expertise, the model does not magically create it. It can smooth the path, but it cannot supply the missing map. It can help someone write a decent email, but not necessarily teach them judgment about what to say, when to say it, or what not to say. It can generate a workable strategy, but not necessarily the intuition needed to know whether the strategy is nonsense.

That limitation matters because it reveals something uncomfortable: tools do not eliminate the need for apprenticeship, they make its absence more visible. A novice with AI can appear competent longer than a novice without AI. But appearance is not mastery. The machine can imitate the outward shape of expertise while leaving the user unprepared for the moments when the script breaks.

Think of a piano student using a playback app. The app may help them reproduce a piece beautifully, but if they cannot hear tension, anticipate phrasing, or recover from a mistake, they are not yet musicians in the full sense. They are operators of a tool. The same is true in knowledge work. AI can shorten the path from ignorance to usefulness, but usefulness is not the same as wisdom.

This is why organizations are tempted by AI in the wrong way. They imagine it can substitute for training. In reality, it often raises the premium on training. The better the tool, the more dangerous it is to use it without judgment. When low competence meets high leverage, errors do not stay small for long.

Why Formation, Not Information, Is the Real Scarce Resource

The deepest connection between family life and AI is this: both reveal that human formation is the bottleneck.

We often talk as though the scarcest resource in modern society is information. It is not. Information is abundant. Advice is abundant. Content is abundant. The scarce resource is the ability to absorb, interpret, embody, and transmit what matters. That ability comes from repeated exposure, guided practice, and a stable set of norms that make learning cumulative.

This is why older generations can look at younger generations and feel bewildered. It is not simply that values changed. The whole process of transmission changed. Parents who had grown up in thickly inherited worlds often assumed their children would inherit by osmosis. But osmosis does not work once the membrane has been punctured by distraction, mobility, and disconnection.

The same dynamic shows up in workplaces. A company can buy software, hire consultants, and deploy AI, yet still fail to produce real expertise if it no longer knows how to apprentice people. Employees can learn isolated tricks without learning the deeper grammar of the field. They can optimize outputs while losing the capacity to recognize what a good output even is.

This is the paradox of abundance. The more outputs we can generate on demand, the easier it becomes to neglect the slow, invisible work that makes outputs trustworthy. But trustworthiness is not produced by speed. It is produced by formed judgment.

Modern systems are increasingly optimized for performance, while the human beings inside them are starving for formation.

That is why so many institutions feel brittle. They can still function, sometimes impressively, but they are hollowed out at the level that matters most. The church can have services without discipleship. The school can have content without character. The workplace can have productivity without mastery. The tools keep improving, while the human pipeline quietly deteriorates.

A Better Model: The Three Layers of Competence

To understand what is happening, it helps to use a simple framework: competence has three layers.

1. Exposure

This is contact with material, ideas, language, or practices. A child hears prayers. A new employee sees a dashboard. A beginner asks an AI to draft a proposal. Exposure is necessary, but it is only the first layer.

2. Repetition

This is repeated practice under some form of guidance. The child joins rituals, the employee works through real cases, the novice edits AI output against feedback. Repetition creates familiarity, and familiarity reduces friction.

3. Judgment

This is the ability to know what matters, what is missing, what is risky, and what is true enough to trust. Judgment comes from reflection, correction, and lived consequences. It is the layer that cannot be outsourced, because it is what tells you whether the tool, tradition, or tactic is actually serving the goal.

Most modern failures happen when we mistake layer 1 for layer 3.

A family assumes exposure to religion will produce faith. A manager assumes exposure to AI will produce competence. A school assumes exposure to concepts will produce understanding. But each layer depends on the previous one, and the jump from repetition to judgment is where most of the real work happens.

This model also clarifies why some uses of AI are genuinely powerful. If someone already has strong judgment, AI can accelerate exposure and repetition. It can free experts from routine and let them focus on higher-order work. But if judgment is missing, the same tool can create the illusion of advancement without the substance.

So the question is not whether AI helps. It does. The question is whether we have a human formation system robust enough to absorb that help without becoming dependent on appearances.

What Strong Transmission Actually Looks Like

Strong transmission is not passive. It is not a speech, a value statement, or a yearly program. It is a pattern of life that makes certain habits normal and certain standards nonnegotiable.

In a religious family, that might mean regular participation, shared rituals, honest conversations, and a home where belief is woven into daily life rather than performed only on special occasions. In a workplace, it might mean real apprenticeship, shadowing, feedback loops, and explicit standards for good work rather than assuming people will figure it out from prompts and dashboards.

In both cases, the same principle applies: people learn what matters by seeing what is repeated, corrected, and protected.

Consider the difference between a child who occasionally hears a parent say faith matters and a child who sees that faith shapes calendar choices, conflict resolution, generosity, and how Sundays are treated. Or the difference between a junior analyst who is handed an AI tool and one who is taught how senior people evaluate uncertainty, recognize patterns, and reject persuasive nonsense.

One experience produces familiarity. The other produces identity.

That distinction matters because many institutions are settling for familiarity and calling it formation. But familiarity without identity is fragile. The moment pressure rises, the person reverts to whatever other system has formed them more deeply.

This is why secularization often begins at home, and why AI failures often begin with novices. In both cases, the problem is not merely loss of information. It is loss of embedded apprenticeship.

Key Takeaways

Do not confuse access with mastery. A child who hears religious language, or an employee who uses AI, is not yet formed. Ask what habits, repetitions, and judgments are actually being built.
Audit your transmission system, not just your message. If you care about values, skills, or standards, look at the daily structure that carries them. The medium of formation matters as much as the content.
Use AI to amplify expertise, not replace apprenticeship. The best use of AI is often to speed up work for people who already have a strong base. It should widen the reach of judgment, not substitute for it.
Protect the environments where judgment is learned. That means time for reflection, correction, and real-world consequences. Judgment cannot be downloaded. It is trained through feedback.
Measure what survives without the tool. If a person, family, or team cannot function at all when the prompt, program, or platform disappears, then the real competence was never internalized.

The Real Question Is Not What We Can Produce, but What We Can Pass On

We often treat religious decline and technological acceleration as separate stories. One belongs to culture, the other to innovation. But together they reveal a single structural truth: modern society is exceptionally good at generating outputs and increasingly bad at forming people who can sustain them.

That is the crisis beneath the crisis. We can produce answers faster than ever, but we struggle to produce the kinds of adults, workers, and communities that know what to do with answers. We can simulate competence, but simulation is not succession.

The deepest challenge, then, is not whether the next generation can access what we had. It is whether they are being shaped into people who can recognize, carry, and improve it. A faith that survives only as nostalgia, and a profession that survives only as software-assisted output, are both signs of the same loss.

The future will belong not to the groups with the most information, nor even the best tools, but to those who remember this simple truth: human beings do not inherit excellence by default. They inherit it through formation.

And once formation weakens, every institution, whether sacred or secular, begins to age faster than it realizes.