AI and the Vanishing Ladder: Rethinking Work in the Age of Automation

The conversation around AI and jobs tends to swing between extremes: utopia or apocalypse. On one side, it promises endless productivity and human flourishing. On the other, it threatens to wipe out entire swaths of the workforce. But the truth is more complicated—and more urgent.

Jobs Aren’t Dying. They’re Shifting.

AI isn’t killing jobs wholesale; it’s reshaping how work gets done. We’re not witnessing a one-for-one replacement of humans by machines. We’re watching workflows being rewritten—with AI as a tireless collaborator.

Still, displacement is real. Repetitive, rules-based jobs are most at risk: data entry, junior analysis, customer service. We've seen this before—the ATM didn’t kill banking jobs; it changed them. The spreadsheet didn’t destroy finance teams; it leveled them up. But today’s shift is faster and deeper. AI doesn’t just optimize a role—it can absorb it entirely.

The Disappearing Rungs

The deeper issue isn’t just that jobs are changing—it’s which jobs are disappearing. We're not just automating roles. We’re automating the bottom rungs of the professional ladder.

Entry-level work has always been where talent grows. Junior developers learn by debugging messy code. Analysts grow by wrestling with raw data. Assistants pick up the rhythms of leadership by osmosis. These roles were never glamorous—but they built careers.

Now?

• McKinsey is training AI to replace junior analysts.
• Tools like Cursor can refactor codebases with a prompt.
• LLM-based agents summarize meetings, manage schedules, and draft content.

This isn’t just disruption. It’s disintermediation of learning. If the junior tier is hollowed out, where will the next generation build their intuition, context, and judgment?

Education Is Outpaced

Universities lag. Bootcamps oversell. Most education systems can't keep up with how fast the ground is shifting. Worse, they still train people for jobs that either won’t exist—or that AI will do better and cheaper.

What we need isn’t more certificates.
It’s more onramps that still exist.

The Barbell Economy

We’re already seeing examples in the real world. Klarna cut hundreds of employees after deploying AI agents to handle tasks once assigned to operations and support teams. McKinsey’s AI initiatives aim to automate junior analyst work across industries. At the same time, companies like PwC and EY are quietly restructuring entire workflows to lean more on AI-enhanced knowledge work, reducing the need for manual white-collar labor across departments. Meanwhile, JPMorgan has deployed AI copilots to assist financial analysts and compliance teams—automating the grunt work, yes, but also scaffolding new roles focused on risk escalation and model validation. In manufacturing, Bosch has begun blending AI with apprenticeship programs—using machine learning to assist junior operators while retaining human oversight.

We’re heading toward a barbell-shaped labor market:

• On one end: elite operators using AI to 10x their output.
• On the other: workers whose jobs are quietly automated away.
• In the middle: a growing void.

No gradual climb. No steady skill progression. Just a jump—or a fall. The middle is soggy—bloated with cost, resistant to change, propped up by vague productivity and busywork. AI is coming not just for the rote and junior, but for the middling roles that survive by navigating bureaucracy and managing inertia.

Manual labor isn’t limited to blue-collar work. Many white-collar roles persist by grinding through approvals, formatting docs, or copying status updates. AI is exposing the inefficiency of this "knowledge work factory floor"—and many of those workflows are being redesigned out of existence.

We risk building a caste system of work:

• Those who command AI.
• Those who compete with it.
• And those who are simply displaced by it.

But what about the counter-narrative? Some argue that AI creates new rungs—that prompt engineer, AI trainer, and model auditor are today’s entry points. While that’s partially true, these roles are often inaccessible to those without prior experience or domain fluency. They're not new rungs so much as entirely different ladders.

Without intervention, AI won’t just concentrate wealth. It will concentrate opportunity. Those with early access, support, and context will surge ahead—not because they’re smarter, but because the ladder was still there for them.

We risk building a caste system of work:

• Those who command AI.
• Those who compete with it.
• And those who are simply displaced by it.

Tacit Knowledge Is At Risk

When a junior product analyst shadowed a customer support escalation, they didn’t just take notes—they learned how to diffuse tension, translate ambiguity into insight, and watch a veteran connect data to action in real time. That kind of tacit knowledge isn’t taught. It’s absorbed.

It’s not just skills that vanish when junior roles disappear—it’s tacit knowledge. The subtle judgment calls, the unwritten norms, the ability to read the room or handle a client call after watching a mentor do it a dozen times. These are things that aren't in the handbook and don’t show up in documentation. They’re absorbed through exposure.

If junior workers aren’t around to observe, contribute, and make small mistakes, they don’t build the context that makes for good decision-making later. As a result, organizations risk growing brittle—optimized for output, but shallow in depth.

Without fresh talent growing through proximity and practice, succession plans stall. Mid-levels become overburdened. Senior talent retires or churns, and no one is ready to step up. The organizational memory shrinks.

Organizations Are Not Ready

Most companies still reflect a pre-AI world. Job titles, performance metrics, and hiring pipelines all assume a talent structure that’s eroding beneath them.

If we eliminate the bottom:

• Who trains the top?
• Where does institutional memory come from?
• How do we promote people who never got to learn?

A Call to Action for Leaders

The easy path is clear: automate the low end, cut costs, and report higher margins.

The harder—and more strategic—path is to rebuild the ladder while embracing AI.

Here’s what that looks like:

• Protect early-stage work
  Don’t just automate it—scaffold it. Assign humans to shadow AI-driven workflows and annotate outcomes. Expect more than passive observation: treat these as active training pipelines. Let people contribute, challenge, and escalate decisions. Build systems where early-stage workers aren’t sidelined—they’re engaged, learning through structured exposure and feedback.
• Rethink performance metrics
  Stop rewarding keystrokes. Start measuring judgment, clarity, and outcomes.
• Build internal training paths
  If the world can’t teach them, you must. Move beyond lip service—build structured apprenticeships where juniors pair with AI copilots and real teams. Give them ownership over specific deliverables, not sandboxed toy projects. Rotate them across domains. Expect real output, real judgment, and real mistakes. That's how growth happens.
• Embrace orchestration, not just execution
  The most valuable employees won’t be the ones who outwork the machine. They’ll be the ones who direct it with clarity and purpose.
• Widen access
  Make AI tools and training available across your org. Don’t let AI fluency become a gatekeeping mechanism. Uplift and support departmental champions—workflow insights and adjustments will come from those closest to the work.

The Future Is AI-Augmented, Not AI-Dominated

The best workers won’t be those who do what AI can’t.
They’ll be the ones who know how to use it wisely.

And the best organizations won’t be those who cut first.
They’ll be the ones who invest in learning, design for hybrid teams, and build a future that includes everyone—not just the AI-native elite.

This isn’t just a time to shrink the workforce.
It’s a time to reshape it.

Final Thought

This isn’t a crisis of efficiency—it’s a crisis of opportunity.
The real test? Creating an AI‑infused workforce built on judgment, escalation, and human oversight.
Leaders who commit to preserving learning spaces while embracing AI will not only win in output—they’ll win sustainably, with people still climbing the ladder, not falling through it.