Artificial intelligence is reducing the number of entry-level office jobs at a moment when young graduates are already struggling to find stable work, raising concerns that companies may be damaging their own long-term talent pipelines.
Researchers at the Federal Reserve Bank of Atlanta have revived a decades-old economic theory from Nobel Prize-winning economist Kenneth Arrow to argue that early-career work is not simply administrative overhead, but a critical training ground for future managers and specialists. Their warning comes as unemployment among young college graduates has climbed above the broader national rate, reversing a pattern that had largely favoured degree holders for years.
The debate over AI entry-level jobs has become more urgent in 2026 as companies across consulting, finance, legal services, marketing and technology increasingly use generative AI tools to handle routine analysis, drafting and administrative work once assigned to junior staff.
Kenneth Arrow’s Theory Returns to the Spotlight
Kenneth Arrow, who later won the Nobel Prize in Economics, argued in a 1962 paper that workers improve primarily through experience gained while performing tasks. The Atlanta Fed researchers applied that principle to modern white-collar employment, concluding that repetitive junior work often serves as the foundation for higher-level expertise later in a career.
According to the researchers, entry-level positions function as a form of applied education that universities cannot fully replicate. Tasks that appear routine today often help younger employees build judgment, institutional knowledge and problem-solving skills that companies later rely on in leadership roles.
The paper argues that removing too many junior positions through automation could eventually weaken firms from within. While businesses may benefit from lower labour costs in the short term, they risk creating a future shortage of experienced workers capable of managing operations, mentoring staff and driving innovation.
The researchers wrote that the activities assigned to junior workers are effectively “the curriculum” through which future productivity is built. If those opportunities disappear, firms may save money today while sacrificing the quality of future leadership.
The warning arrives as hiring activity across white-collar industries remains subdued. Many employers expanded aggressively after the pandemic before reducing headcount over the past two years amid slower economic growth, geopolitical tensions and uncertainty linked to tariffs and the war in Iran.
Young Graduates Face a Tougher Labour Market
The cooling job market has hit younger workers particularly hard. In some sectors, recent graduates are experiencing unemployment rates close to those seen among peers without university degrees, adding pressure to the value proposition of higher education.
A growing number of Gen Z workers are also reconsidering traditional office careers in favour of skilled trades and vocational work, where labour shortages and wage growth have created more stable opportunities.
Research from McKinsey published earlier this year estimated that generative AI could automate a significant share of routine cognitive tasks across professional services, particularly work involving document preparation, scheduling and basic research. While many economists expect AI to create new categories of employment over time, the transition period may leave a gap in workforce development if companies sharply reduce hiring at the bottom of the corporate ladder.
That challenge extends beyond individual firms. Arrow’s original theory suggested that learning-by-doing contributes to productivity gains across entire industries, not just within one company. The Atlanta Fed researchers argue this means widespread automation of junior tasks could eventually weaken broader economic growth if fewer workers gain the experience needed to move into senior positions.
Comparable debates emerged during earlier waves of industrial automation in manufacturing, when companies reduced apprenticeship-style roles only to later confront shortages of experienced technical staff. Economists have increasingly warned that AI may create a similar dynamic in professional services.
Why Companies May Still Push Ahead With Automation
Despite the risks outlined in the research, businesses still have strong financial incentives to automate. AI systems can complete some administrative and analytical tasks at far lower cost than hiring and training junior employees, particularly as firms remain under pressure to protect profit margins.
To counterbalance those incentives, the Atlanta Fed researchers proposed policies that would encourage companies to continue employing younger workers. Their suggestions included taxes tied to automation-driven profits alongside subsidies for firms that expand opportunities for entry-level employees.
The researchers argued that such measures could help preserve learning opportunities while still allowing businesses to benefit from productivity gains linked to AI.
For now, the immediate financial impact of automation appears to favour employers rather than workers. The paper concludes that if companies collectively reduce learning opportunities for younger employees, much of the economic cost will initially fall on workers through lower wages, weaker career progression and fewer pathways into professional management roles.
As AI adoption accelerates, policymakers and corporate leaders may soon face a broader question than whether automation cuts costs. They may also need to decide how future expertise is developed when fewer workers are given the chance to learn on the job.
