National Apprenticeship Week should do more than celebrate a promising workforce pathway. In an age of artificial intelligence, it should draw attention to a deeper problem. America is losing one of the main ways people become good at work in the first place.
For generations, the pathway to workforce expertise was straightforward. People start with smaller tasks, repeat them, improve, and gradually assume more responsibility. In 1936, aeronautical engineer Theodore P. Wright gave this process an economic logic.
He found that every time cumulative output doubled in airline production, the cost of building each plane dropped by about 20%. Workers learned by doing, and that learning compounded over time.
The result was dubbed Wright’s Law, or the learning curve.
And what he observed in factories described something larger that explained how industries advance. Experience is an asset, not only a work byproduct. Firms get better because people and organizations improve through repetition, feedback, and practice. The more we do something, the better and more efficient we are at doing it.
That truth still matters. But something fundamental is changing.
Artificial intelligence is automating many entry-level tasks that once allowed novice workers to learn. The lower rungs of the career ladder, where people made mistakes, developed judgment, and built confidence, are weakening.
A recent Wall Street Journal commentary shows that AI is “destroying the old learning curve” because machines can reproduce forms of expertise that took humans years to acquire. What required time, supervision, and practice is now simulated instantly by software trained on massive amounts of data.
That creates a paradox at the center of the economy. Just as demand for higher-order skills is rising, the opportunities to develop those skills are shrinking.
The real problem isn’t only whether AI eliminates jobs. It’s whether AI erodes the career pathways through which people become capable enough to do better jobs later. The issue isn’t only employment. It’s the weakening of the learning process itself.
That problem is especially visible in knowledge work. Entry-level assignments that once trained young professionals, such as drafting memos, summarizing reports, doing preliminary analysis, and building presentations, are increasingly handled by AI tools.
As Matt Sigelman of Burning Glass Institute and his colleagues argue, AI changes the value of expertise. It does this not by eliminating the need for knowledge but by changing where knowledge lives. Once, expertise was accumulated in people over time. Now, more of it is embedded in tools.
That makes human development harder, not less important.
MIT economist David Autor sharpens the point. AI is especially good at information acquisition and routine cognitive tasks. But it complements human judgment, tacit skill, and social coordination. The most valuable workers don’t simply know facts. They know how to interpret context, exercise discretion, collaborate with others, and make sound decisions in ambiguous situations.
But those capacities don’t appear automatically. They’re built through practice.
That’s why earn-and-learn apprenticeships matter so much. They aren’t merely another workforce option. They’re one of the few remaining systems deliberately designed to preserve learning by doing.
At their best, apprenticeships combine paid work, structured learning, mentoring, feedback, and increasing responsibility. They allow people to develop skills in real settings while earning wages and building confidence. They do more than connect people to jobs. They rebuild the human learning curve that AI is flattening.
That’s why apprenticeships may be the new Wright’s Law for human capital.
Wright showed that each doubling of cumulative output lowered production costs because workers and firms got better through repetition. The modern apprenticeship equivalent is that each doubling of meaningful experience lowers the cost of competence. Learners become more productive, capable, and adaptable. What cumulative production did for factories, cumulative experience does for people.
This principle extends beyond traditional registered apprenticeships. It includes youth apprenticeships, internships with skill progression, clinical rotations, dual-enrollment pathways connected to work, and project-based learning tied to external audiences. The common thread is not the label. It’s structured practice. These models create opportunities for learners to apply knowledge, receive guidance, and improve under real-world conditions.
Economic historian Joel Mokyr explains why this matters. He shows that apprenticeships were central to Europe’s industrial rise because they transmitted tacit knowledge from masters to novices. They weren’t just training arrangements.
They were institutions for spreading practical know-how, disciplined experimentation, and habits of improvement. Long before economists formalized learning curves, apprenticeships were helping societies move down them.
That older lesson has new relevance. The digital economy doesn’t suffer from a shortage of information. It suffers from a shortage of experience transmission.
A worker can ask AI for a summary, a draft, a recommendation, or a block of code. But AI can’t fully supply the human process of learning through responsibility, error, correction, and judgment. It can’t replace what happens when a person must answer to a supervisor, serve a customer, solve a problem with others, or make a decision with consequences. Those aren’t abstract forms of knowledge. They’re living ones.
As AI lowers the cost of accessing information, it raises the premium on interpretation. In a world where “knowing that” becomes cheaper, “knowing how” becomes more valuable. That’s why apprenticeships are an essential infrastructure for an AI economy.
The policy question is not whether the country can preserve every entry-level task that software can automate. It can’t. The better question is whether America can build enough structured pathways for people to acquire judgment, tacit skill, and confidence anyway. If the old learning curve is being broken by machines, then the country needs institutions that deliberately rebuild it.
That’s the larger significance of National Apprenticeship Week. It shouldn’t only celebrate apprenticeship as an alternative pathway for some students or workers. It should present apprenticeship as a central strategy for preserving human development in a labor market increasingly shaped by machines.
Future prosperity will belong not only to those who automate intelligence, but also to those who organize experience. Apprenticeships do that. They translate information into capability. They connect wages to learning. They bind education to responsibility. And they preserve the human truth at the heart of Wright’s Law. We get better by doing.
In that sense, earn-and-learn apprenticeships are more than a workforce strategy. They are the new Wright’s Law.
Bruno V. Manno is a senior adviser at the Progressive Policy Institute and leads its Pathways to Opportunity What Works Lab.