Who Gets to Start?

AI is not our best friend when it comes to unemployment.

A global survey by the Oliver Wyman Forum found that the share of CEOs planning to shift away from junior roles over the next two years had more than doubled, rising from 17% in 2025 to 43% in 2026.

An analysis by Goldman Sachs Research estimated that AI reduced monthly payroll growth in the United States by roughly 16,000 jobs during the previous year and raised the unemployment rate by 0.1 percentage point. Goldman Sachs also found that AI-assisted growth created jobs in other areas, offsetting some of those losses, and cautioned that its estimate might not capture every indirect employment gain.

The overall effect on the labor market may still appear modest, but the pressure does not seem to be distributed evenly. Goldman Sachs concluded that the negative effects of AI on employment were falling largely on younger and less-experienced workers.

Treating the disappearance of junior work as an inevitable consequence of technological progress obscures a more important question: If companies automate the work through which young employees once learned, who will be allowed to become senior employees in the first place?

The Work Was Also Training

The reason for replacing junior tasks appears simple: AI can operate quickly and at a massive scale.

Artificial intelligence can automate foundational tasks such as drafting documents, scheduling meetings, cleaning data, summarizing reports, debugging basic code, copy editing, and formatting spreadsheets. These are often described as “busy work,” which makes their disappearance sound harmless.

However, these tasks were not valuable only because someone needed to perform them. They also served as training for inexperienced employees.

A junior worker who researches a report begins to understand which sources are reliable. Someone who takes notes during a meeting learns how senior employees negotiate and make decisions. A new programmer who debugs simple code gradually develops the judgment required to solve more complicated problems.

The work may be repetitive, but the learning is not.

This is where the career ladder comes into play. An intern becomes a junior employee, then an associate, manager, and perhaps eventually a member of the senior leadership team. These positions may have significantly different responsibilities, but they are connected.

The skills learned in one role become the foundation for the next. Like the parts of a machine, an employee learns, adapts, and grows, with each experience contributing to the next stage of a career.

AI may eliminate tasks at the bottom without eliminating the need for people at the top. If it removes the work through which inexperienced employees gain experience, where are experienced workers supposed to come from?

The Experience Paradox

The result is an experience paradox.

Companies increasingly expect entry-level applicants to arrive with AI literacy, professional judgment, strategic thinking, communication skills, and industry experience. According to the National Association of Colleges and Employers, more than one-third of entry-level jobs represented in its 2026 employer survey required AI skills, nearly three times the share reported only months earlier.

But how does a 22-year-old acquire professional judgment and industry experience without first being allowed into the workplace?

This contradiction has produced the increasingly absurd phenomenon of supposedly “entry-level” jobs asking applicants to have several years of previous experience. Indeed notes that many entry-level listings request two to five years of experience, even though employers may still consider applicants who do not meet every preferred qualification.

AI could intensify this mismatch. Employers may use the technology to eliminate the work traditionally performed by beginners, then expect those same beginners to apply with skills that were once developed through that work.

In other words, AI may remove the mechanism through which people become qualified for the jobs that remain.

When Experience Becomes a Privilege

But who can afford to gain experience outside a paid entry-level job?

Imagine two equally capable graduates.

Abigail comes from a wealthy family. She has an Ivy League degree, professional connections, and access to expensive AI and software training. She can afford to accept an unpaid internship and spend six months networking without earning a stable income.

Bob comes from a family with limited financial resources and needs an income immediately. He cannot afford to work without pay or spend months developing connections while waiting for the right opportunity.

If traditional entry-level positions disappear, Abigail may still find informal pathways into professional life. Bob may not.

The disappearance of entry-level work does not necessarily create a world without pathways into professional careers. Instead, it risks creating a world in which those pathways become private.

Experience becomes something purchased through family wealth, personal connections, prestigious universities, unpaid internships, and access to specialized training. The career ladder still exists, but only certain people are given access to its first rung.

That makes the future of entry-level work more than a technological issue. It also makes it a question of economic mobility.

Automation Is a Choice

At the end of the day, replacing workers with AI is not always inevitable. In many cases, it is an institutional decision.

Artificial intelligence can perform many tasks, but employers still decide which parts of a job it will perform and how its productivity gains will be used.

A company can use AI to eliminate junior positions and reduce labor costs. It can also use AI to remove repetitive work while allowing junior employees to focus on analysis, collaboration, and decision-making under the guidance of experienced workers.

Employers could use productivity gains to shorten workweeks, expand mentorship programs, improve training, or allow young employees to take on more meaningful responsibilities. They could redesign entry-level work rather than eliminate it.

Even the Oliver Wyman survey complicates the idea that every advanced company is abandoning junior talent. Among companies reporting strong returns from AI, some were moving toward hiring more junior workers because they saw value in digitally fluent, AI-capable employees.

It is a technological fact that AI can perform many junior-level tasks. Whether that ability is used to eliminate workers or improve their training remains a political and economic choice made by companies and governments.

The real question is whether employers will cut pathways into professional life or pay humans to learn, supervise, question, and improve the work AI produces.

Do Not Eliminate Apprenticeship

Although AI is undeniably a powerful force, humans still determine its role and influence in society.

In a future where people recognize both AI’s capabilities and the importance of human workers, the technology could eliminate drudgery without eliminating apprenticeship.

Junior employees do not need to spend their careers formatting spreadsheets or scheduling meetings. But they do need opportunities to observe, practice, make mistakes, receive feedback, and develop judgment.

Entry-level work should evolve alongside AI. It should not disappear because of it.

The greatest danger of AI may not be that it eliminates the bottom rung of the professional ladder.

It is that we continue demanding that young workers somehow arrive at the second rung anyway.