AI, Apprenticeship, and the Future of Learning

On my first day as a structural engineer in the aerospace industry, my supervisor walked me into a room roughly the size of a football field. Every desk was filled with engineers. Most of them had master’s degrees or more.

Then he gave me a piece of advice that stayed with me for the rest of my career.

Despite my brand new master’s degree in structural engineering from Stanford, I was now the lowest-ranking person in the room. The expectation was simple: watch, listen, and learn by doing the work.

The early assignments were not glamorous. They involved checking calculations, reviewing drawings, and fixing problems that more experienced engineers had already seen many times before. In fact, my very first job was to design a nut to fasten things to the Saturn V second stage rocket. A nut – not very glamorous.

It was slow work. Sometimes tedious work.

But that was where the learning happened.

Those first years functioned as an informal apprenticeship. By repeating small tasks, you gradually developed intuition. You learned where mistakes usually occurred. You began to recognize patterns before problems were fully explained. You listened to debates between experienced engineers and slowly absorbed the reasoning behind their decisions.

Over time the same process repeated itself across many professions.

Junior analysts cleansd messy data.
Junior programmers fixed bugs.
Junior lawyers reviewed endless stacks of documents.
Junior accountants reconciled spreadsheets.
Junior doctors performed routine examinations under supervision.

From the outside those tasks often looked inefficient. A senior employee could finish them faster.

But efficiency was never the point.

Exposure was.

Those early tasks formed the bottom rung of the career ladder — the place where practical knowledge accumulated.

And that bottom rung may be quietly disappearing.

If artificial intelligence removes many of the small tasks where that repetition once occurred, we face an interesting question.

Where will the next generation develop that intuition and gain knowledge that evolves into wisdom?

Some industries may not notice the change immediately. The pipeline of experienced professionals built under the older apprenticeship system will continue working for years. Their accumulated knowledge will carry organizations forward for a while.

But over time a gap could appear.

Senior experts may still exist.

AI tools may still exist.

But the middle layer of developing professionals — the people quietly building the next generation of expertise — may become thinner.

In effect, the training pipeline could narrow.

That possibility raises an uncomfortable question for many professions.

If the bottom rung of the ladder disappears, how do people climb the ladder at all?

At the moment there is no single answer. But a few possibilities are beginning to emerge.

In the short term, AI may simply become part of the apprenticeship itself. New professionals may spend less time doing repetitive work and more time learning how to supervise and interpret AI-generated results.

Instead of manually reviewing thousands of data points, a junior analyst might review the AI’s conclusions and learn how to identify subtle errors.

Instead of writing every line of code from scratch, a new programmer might evaluate AI-generated code and understand why certain solutions succeed while others fail.

“The early tasks where people once learned their professions are slowly disappearing.”

In other words, the apprenticeship may shift from producing work to evaluating work — or testing it carefully before it ever reaches the real world.

Some companies may create more structured mentoring programs where experienced professionals intentionally guide younger workers through complex problems rather than simply assigning small tasks. Or may guide them through thorough testing to uncover problems or mistakes and why they are not suitable and how they occurred.

That might accelerate certain types of learning, but it could also create blind spots if people never experience the underlying mechanics themselves.

Longer term, industries may need to rethink how professional knowledge is passed down.

Educational institutions may also need to adapt. Universities have traditionally focused on theoretical knowledge while expecting workplaces to provide practical training. If workplaces begin eliminating entry-level tasks, schools may need to incorporate more real-world problem solving into their programs.

Simulation may also play a role. Fields such as aviation and medicine already use sophisticated training simulations to replicate real scenarios. Similar approaches could emerge in engineering, law, finance, and software development.

But even with those adjustments, one truth remains.

Experience still matters.

The difference between someone who has seen a hundred problems and someone who has seen ten is often enormous.

That kind of accumulated pattern recognition is difficult to replicate artificially.

Which brings us back to the larger question raised by artificial intelligence.

AI can dramatically increase productivity. It can remove tedious work, accelerate research, and help experienced professionals solve problems faster.

But if we remove too many of the small learning steps along the way, we may accidentally weaken the pipeline that produces the next generation of expertise.

The challenge is not stopping technological progress. That has never worked and probably never will.

The challenge is making sure the ladder of learning still exists — even if the rungs have to be rebuilt in a different form.

For generations, professional knowledge grew one small problem at a time.

Artificial intelligence may change how those problems are solved. But it cannot eliminate the need for people to learn how to think through them.

If the bottom rung of the ladder disappears, we will have to invent a new one.
Because expertise has never appeared fully formed. It grows slowly — through small problems, repeated exposure, and the quiet accumulation of experience.

Every expert you have ever met started out doing simple things that didn’t look very important at the time. Sometimes even designing a nut for a rocket.