Automation, AI, and the Intergenerational Transmission of Knowledge

Recent advances in Artificial Intelligence (AI) have sparked expectations of unprecedented productivity growth. However, by enabling senior workers to accomplish more tasks independently, AI may inadvertently reduce entry-level opportunities, raising concerns about how future generations will acquire essential expertise. This paper develops a model to examine how advanced automation affects the intergenerational transmission of knowledge. The analysis uncovers a critical trade-off: automating entry-level tasks yields immediate productivity gains but risks undermining long-term economic growth by eroding younger workers' acquisition of tacit skills. Back-of-the-envelope calculations suggest that AI-driven entry-level automation could lower the long-run annual growth rate of U.S. per capita output by 0.05 to 0.35 percentage points, depending on the scale of automation. I further demonstrate that AI co-pilots -- systems providing scalable access to tacit-like expertise previously acquired only through direct experience -- can partially mitigate these adverse effects by assisting individuals who fail to develop adequate skills early in their careers. However, co-pilots are not always beneficial, as they may also weaken the incentives of junior workers to engage in hands-on learning. These findings challenge the optimistic view that AI will automatically sustain productivity growth, highlighting instead the importance of safeguarding or actively creating new entry-level opportunities to fully unlock AI's potential.