Are Apprentices an Endangered Species?

The experience of apprentices and trainees in the real world is not that different from Luke’s. Take, for example, the junior lawyer drowning under a pile of paperwork, or the corporate intern taking coffee orders, or a line of aspiring chefs shucking oysters all day. As grueling as it might seem, this scut work is often viewed as a necessary evil because it’s the only way many apprentices are able to compensate their masters for training them. 

“The master might say, ‘Okay, I’m going to train you. But in return, I’m going to have you working for me for a while, and I’m going to pay you less than how much you produce,’” says Luis Rayo, a professor of strategy at the Kellogg School and an expert on contract theory and organizational economics.    

But the recent rise of AI has complicated this long-standing dynamic. Large-language models like Claude and Gemini are performing many common apprentice tasks increasingly well, often with much greater efficiency than the average trainee. Indeed, recent research already shows a decline in hiring for entry-level, but not experienced, workers in the occupations most exposed to generative AI.   

“Apprenticeship-like systems are a very common way in which economies solve the difficult problem of transferring human capital from one generation to the next,” Rayo says. “AI risks destroying that system, and the problem then is, who’s going to train the next generation of experts?”  

To better understand the possibility of that happening, Rayo collaborated with Luis Garicano of the London School of Economics to mathematically model what happens when AI enters the apprentice–master relationship.   

They found that the more tasks AI takes over from the apprentice, the less viable apprenticeships become. Yet they also found that AI could help elevate the peak ability and productivity of advanced-level apprentices, which could increase the value of apprentices and give masters more incentive to hire one.   

“So it’s a race between the two [effects of AI],” Rayo says. “One shrinks profits, the other one grows profits, and whichever one wins determines whether an apprenticeship is profitable.”   

Here comes AI 

Rayo’s interest in this topic stems from his prior research, in which he and his coauthors modeled the traditional apprentice–master relationship. The model has two baseline assumptions: the apprentice does not have enough money to pay the master for the training and cannot simply promise to pay later. Thus, the master wants to stretch out the training period to get the most value from the apprentice while training takes place. 

Rayo and Garicano, both fellows at the Centre for Economic Policy Research, identified two main effects after introducing AI into this model.   

First, AI raises the floor for what apprentices need to do to provide value. Since AI tools can complete many of the apprentice’s basic tasks almost entirely on their own, the apprentice is no longer needed for menial work. As a result, the apprentice is less valuable to the master.    

“Now that AI does that [work] essentially for free, the currency that apprentices are using to buy knowledge is vanishing,” Rayo says. 

Second, AI raises the ceiling for what an advanced apprentice can offer. AI tools can complement some of the higher-level work that trainees do, which cranks up the value of their work while giving them so much more to learn.   

“So, for high-enough knowledge levels for the apprentice, AI is good for you. It complements your skills; it makes you more productive,” Rayo says. “And if the ceiling becomes super high, then there’s still a lot of productivity growth potential.”   

A race for the ages   

The economists use the setting of a law firm to illustrate how this might play out in real life. With AI, a junior associate—the apprentice—no longer has as much menial work to do, so they are trained to do more advanced work and bill more for it. In other words, the floor is higher. But because junior associates start off doing more advanced work, they are able to finish their training much more quickly than before. So, despite the elevated value they bring at the outset, the shorter apprenticeship ultimately reduces the long-term value they bring to the firm’s partners. So the more AI can do, the less incentive the firm has to take on the costs of hiring and training junior associates.   

At the same time, AI also helps not only senior partners, but also highly trained junior associates work more efficiently and bill more than before. There’s also a lot more that junior associates can absorb and do. If this raises their profitability high enough to cover the cost of hiring and training them, then they become all the more desirable for the firm.    

“The question is, which one is growing faster—the floor or the ceiling?” Rayo says.    

To answer that question for a particular apprenticeship, a company would need to track how much value a highly trained individual produces with the help of AI versus how much value AI creates on its own. According to the model, an apprenticeship is guaranteed to be worth its while if that ratio is greater than Euler’s number “e” (2.71828), the famous mathematical constant widely used in the world of finance.   

A bright future?   

The more the gap between the floor and the ceiling of an apprenticeship widens, the more value the apprenticeship holds. The smaller the gap is, the less value it holds.   

But if the gap between the ceiling and floor shrinks enough, not just for a single apprenticeship but across the workforce, society could face serious consequences.    

“Human knowledge would start disappearing, and then the robots are just going to do whatever they’re able to do on their own, without the added benefit of advanced human knowledge,” Rayo says. “It’s going to be a case where, instead of accumulating knowledge as a society, we start losing knowledge.”   

To address this worst-case scenario and keep apprenticeships alive, businesses could require apprentices or interns to pay the business back after completing their training by agreeing to work for lower pay for an extended period. Or the government could subsidize apprenticeships through grants or loans. But both of these options come at a significant cost, either to the apprentice or society.   

A more-reasonable approach might be to adapt and improve the education system. Universities, for instance, could help prepare future apprentices and interns to enter the workforce at a high-enough level that they immediately bring more value to the company than AI does alone.    

For that to occur, Rayo believes that the education system would need to teach its students extremely valuable knowledge, especially general principles that are broadly applicable to many types of work.    

That’s the kind of information that “gives people the ability to perform tasks that AI cannot do on its own,” he says. “The more fundamental principles you learn from different disciplines, the better position you’re going to be in to direct AI and to judge the quality of AI’s output.”  

These strategies might be less necessary, however, if AI widens the gap between the floor and ceiling for apprenticeships as a whole. In that scenario, apprentices naturally would start out doing higher-skilled work. And they would still have much to learn and offer because of how much AI has increased their productivity and expanded the scope of their work. With a proper education, for example, a young padawan like Luke might have begun his apprenticeship with much more control over his emotions and the Force and ended up developing a greater level of power.  

“We’d spend more time doing advanced things, and our final level of productivity—the ceiling that we reach once we’re fully trained—would just keep growing,” Rayo says. “That’s a bright future with extreme amounts of productivity and with knowledge transmission from one generation to the next.”