AI Is Killing the Career Ladder: Why Young Workers Are the New Canaries in the Coal Mine

AI isn’t just another productivity tool anymore—it’s starting to reshape how careers are built, especially for people at the very beginning of their working lives. New research from labor economists tracking millions of workers suggests that AI is already changing who gets hired, what skills matter, and how risky it is to pick a career path.

Instead of a predictable "career ladder," we may be heading toward a more flexible but uncertain "career lattice"—with more sideways moves, faster reskilling, and a bigger role for AI as a learning companion.

What the Data Says: Young Workers Are Getting Hit First

Economists recently analyzed millions of U.S. workers using payroll data from ADP to understand how AI is affecting jobs. They compared roles that are highly exposed to AI (like software development, customer service, and administrative work) with those that are less exposed (such as transportation or construction).

At first glance, the overall job numbers don’t look dramatic—total employment in AI-exposed jobs hasn’t collapsed. But when you zoom in on young workers, a very different picture appears.

Key finding: In jobs more exposed to AI, young workers are seeing 16% slower employment growth compared to similar workers in less-exposed jobs.

In other words, the people trying to get their foot in the door—new grads, early-career professionals, junior hires—are the ones feeling the squeeze. More experienced workers in the same occupations are still seeing employment growth roughly on trend.

This is why the researchers describe young workers as the "canaries in the coal mine" for AI’s impact on the labor market. Their struggles may be an early signal of deeper, longer-term changes.

Why AI Overlaps So Much With Entry-Level Work

To understand what’s happening, it helps to look at what young workers actually do on the job.

When people first enter a field, most of their work is about implementation—applying what they learned in school, following instructions, writing code from specs, drafting documents, answering routine customer questions, or processing standard requests.

These are exactly the kinds of tasks today’s AI systems are getting good at: structured, well-documented, and based on "book knowledge." If something can be written down in a manual or taught in a class, there’s a decent chance AI can help automate or accelerate it.

By contrast, more experienced workers spend more time on things AI still struggles with, such as:

• Tacit knowledge – Hyper-local context, unwritten rules, and "how things really work" inside a company or industry.
• Strategic thinking – Deciding what should be done, not just how to do it.
• Social interaction – Managing people, navigating politics, building trust, and reading situations.

These skills are built over years on the job, not from textbooks. That gives experienced workers a relative advantage over both AI and younger colleagues whose strengths lie in exactly the areas AI is rapidly improving.

Is This Just a Blip—or a Structural Shift?

One big question is whether this is just a temporary side effect of broader economic conditions, or a deeper, AI-driven shift in how the labor market works.

The researchers tested several alternative explanations:

• Interest rate changes: Sectors like construction and transportation are very sensitive to interest rates but are less exposed to AI. The patterns in the data don’t line up with interest rates being the main driver.
• Tech over-hiring and layoffs: Even when they removed tech companies and computer-related jobs from the analysis, the core results stayed the same.

That doesn’t prove AI is the only cause—there’s no perfect "world with AI vs. world without AI" experiment—but it does suggest we’re seeing the early stages of a structural change, not just a short-term wobble.

And unlike past technologies, AI has another twist: the speed of improvement. The capabilities of modern AI models have jumped dramatically in just a few years. As new types of work emerge, there’s a real question: will humans get those new tasks, or will AI advance fast enough to take them on too?

This same concern shows up in other research on AI’s broader risks and misuse, including how AI can generate misleading or low-quality advice, sometimes called "trend slop" in AI-generated content.

The Career Ladder Is Breaking—Enter the Career Lattice

For most of the 20th century, careers were often described as ladders: you start at the bottom, climb step by step, and eventually reach middle or senior management if you stick with it.

AI is putting pressure on that model in two ways:

• Entry rungs are shakier: If AI can do much of the implementation work, companies may hire fewer juniors and invest less in training them.
• Middle rungs may shrink: If fewer people get early experience, there’s a smaller pipeline of future managers and experts.

From a company’s perspective, it’s rational to hire only as many juniors as they need in the short term. From society’s perspective, that’s a problem: we underinvest in the next generation of skilled workers because any one firm doesn’t capture the full benefit of training someone who might later leave.

That’s where the idea of a career lattice comes in. Instead of a single, narrow ladder, a lattice is a web of paths: sideways moves, reskilling, and switching professions as demand changes.

If AI can truly help people learn faster and more effectively, it could make it much easier to:

• Switch careers midstream when a field declines.
• Move into emerging, high-demand roles more quickly.
• Recover from disruptions like layoffs or industry shifts.

In the best case, AI could turn career risk into career flexibility—if we design education and training systems to take advantage of it.

What AI Still Can’t Do Well (Yet)

Despite rapid progress, there are three big areas where AI is still limited, especially in the short to medium term:

• Physical tasks: Without major advances in robotics, many hands-on jobs remain hard to automate.
• Strategic thinking: Setting goals, making tradeoffs, and deciding what should be done are still very human tasks.
• Social interaction: Building relationships, leading teams, persuading others, and reading subtle cues are deeply human skills.

In many future jobs, work may look like this: AI agents handle much of the implementation, while humans focus on guiding them—deciding what to build, how to prioritize, and what "good" looks like.

That’s essentially a managerial role, even if your title isn’t "manager." The ability to clearly express what you want, evaluate AI output, and steer complex projects will become increasingly valuable.

AI as a Learning Engine, Not Just an Automation Engine

There’s a lot of debate about whether AI will replace workers or augment them. One of the most promising paths for augmentation is using AI as a learning tool.

Historically, the technologies that helped workers the most were the ones that expanded the set of tasks they could do—like education, computers, and the internet. Technologies that purely automate tasks tend to shrink what humans do.

AI can go either way. It can automate away tasks, or it can help people:

• Learn new skills faster.
• Enter fields that used to require years of specialized training.
• Handle a wider range of responsibilities with a smaller team.

Consider a startup founder with a tiny team who uses AI to draft legal documents, prototype designs, write code, and analyze data. That’s a clear example of augmentation: the founder isn’t being replaced—they’re doing more than was previously possible.

Whether you’re being automated or augmented depends on what happens to your task list:

• If AI shrinks what you do, you risk being automated.
• If AI expands what you can do, you’re being augmented.

Education is the classic way to expand human capability. AI-powered, personalized learning could be the biggest shift in how we learn in 100 years—if we build and adopt it thoughtfully. Some platforms are already experimenting with this, like AI tutors that guide you through problems instead of just giving you the answer.

That same idea—AI as a structured assistant rather than a black box—is also showing up in other domains, from math research to AI systems that help scientists generate and test new hypotheses.

What Young Workers Should Focus On Now

If you’re a student or early in your career, the landscape can feel intimidating. But there are concrete steps you can take to position yourself well in an AI-shaped labor market.

1. Use AI Tools—Don’t Ignore Them

Get hands-on with AI as much as possible. Treat it like a calculator for thinking, not a shortcut for avoiding work. Build projects with it, experiment with different tools, and pay attention to:

• What it does well.
• Where it fails or hallucinates.
• How to prompt and guide it effectively.

The more you understand its strengths and weaknesses, the better you’ll be at using it as leverage instead of competition.

2. Develop Strategic Thinking

AI can help you implement, but you still need to decide what to implement and why. That means practicing:

• Breaking vague goals into clear, actionable steps.
• Evaluating tradeoffs and risks.
• Designing processes and systems, not just doing tasks.

Even when using AI for technical work—like checking math proofs or exploring models—it can be smarter to let AI handle the mechanical parts while you focus on understanding the ideas and deciding which direction to pursue.

3. Protect and Grow Your Human Skills

Some skills are likely to become more valuable, not less, in an AI-heavy world:

• Communication: Writing, speaking, and explaining ideas clearly.
• Collaboration: Working with others, managing conflict, and building trust.
• Reflection and judgment: Deciding what you actually want to build or pursue, not just what you can.

These are the skills that help you decide what to delegate to AI—and what to keep as deeply human work.

AI, Inequality, and the Value of Learning

AI could reshape not just jobs, but also inequality and incentives to learn.

Imagine a world where AI makes it easy for almost anyone to produce top-tier work in many fields. If the hardest tasks are handled by AI, the gap between the "best" and the "average" worker might shrink. That could mean lower inequality—but also less payoff for investing heavily in learning traditional skills.

On the other hand, if AI greatly increases the value of strategic thinking, leadership, and social skills, then people who develop those capabilities could become even more valuable. That would strengthen the incentive to learn deeply and think hard, especially in areas AI can’t easily replace.

Which path we end up on will depend on how we design education, how companies choose to deploy AI, and how individuals decide what to delegate and what to keep.

From Fear to Agency: Building a Career in the Age of AI

It’s natural to worry about "staying ahead of AI," especially when you see it solving problems or answering questions you once thought were uniquely human. But that framing can be limiting.

A more useful question is: How do I work with AI to become more capable, more creative, and more adaptable?

That means:

• Leaning into AI as a learning and thinking partner.
• Building skills in strategy, communication, and social interaction.
• Expecting a career lattice, not a straight ladder—and preparing to move sideways when needed.

Underneath all of this is something simple but powerful: your identity, your values, and your sense of what you want to build in the world. AI can help with implementation, but it can’t tell you what matters to you. That part is still entirely human—and it’s where your long-term advantage lies.