Future of work

This article examines how generative AI could reshape organizational hierarchies, particularly by shrinking the number of entry-level and junior roles. If routine early-career tasks are increasingly automated, then rather than operating with a traditional “pyramid” structure, firms may move to one that is more diamond-shaped—narrow at the bottom and top, wider in the middle.  

The early evidence is mixed. Studies by Erik Brynjolfsson and colleagues and by Bouke Klein Teeselink suggest that increasing AI adoption is associated with weaker entry-level hiring and employment outcomes among young workers. However, research by Morgan Frank and co-authors indicates that graduates’ prospects in AI-exposed sectors were deteriorating even before the release of ChatGPT, reflecting broader structural and cyclical pressures rather than AI alone.  

The article argues that entry-level work is especially vulnerable because it often consists of document-heavy, repetitive tasks that large language models can perform more cheaply and efficiently. This raises concerns that junior workers may lose traditional “learning-by-doing” pathways, especially if firms rely on AI or external labour markets to provide early-career training.  

Despite these risks, the author outlines three reasons why firms should resist cutting junior roles. First, the long-run effects of AI on work remain uncertain. Second, eliminating early-career pipelines will create future skill shortages and succession risks. Third, younger workers are typically more comfortable using AI tools and less constrained by legacy work practices, making them valuable for organizational adaptation.  

Evidence from OpenAI suggests that workers aged 18–29 are more than twice as likely to use ChatGPT at work as those over the age of 50, reinforcing the idea that junior staff may be key drivers of AI literacy. Indeed’s head of AI, Hannah Calhoon, emphasizes the strategic value of entry-level hiring as a mechanism for organizational change.