Renewing the Master-Apprentice Model in the Age of AI
AI automates routines. But what happens to the junior who can no longer build intuition through repetition?
From junior to senior. I have experienced this developmental arc myself and seen it hundreds of times. In the beginning, the learning curve is steep: lots of repetition and mistakes. Then something clicks, and you realize you know more and more.
AI offers a tempting opportunity: automate the routines, free up the seniors’ time. I believe this can work. But I have started to wonder if we have all the pieces in place. How do we renew the master-apprentice model in the age of AI?
Researchers from MIT and Stanford have identified a “dual deficit.” When we eliminate the routines through which intuition is built, we might accidentally remove the stepping stones of learning. Highly AI-supported experts achieved results quickly, but their performance dropped in independent decision-making situations.
The research points in two directions:
- “Think-Aloud” Mentoring: A senior explains their thought process out loud: “Why did I choose this? What risks do I see that the data doesn’t show?” This sounds smart and is something we are already doing.
- Progressive Learning Path: Building the stepping stones. The same principle is first learned in a simple environment, then in a more demanding one. This sounds slow, perhaps artificial. But the research shows that organizations operating this way can grow juniors into seniors in three months.
This gives me pause. I have heard of organizations freezing junior recruitment: “AI will handle it; let’s only hire experienced people.” That looks sensible in the short term. But what about in five years, when the next generation of seniors has clocked out?
I believe that organizations that budget for developing expertise as seriously as they do for AI tools are writing the next success story. Conversely, those that do not, are not.
Aspenly · Thinking