Something deeply unsettling is happening in our organizations.
Just look at the numbers from the Brookings Institution's recent research:
53% of entry-level market research work faces AI replacement (vs. 9% for managers)
67% of junior sales tasks are at risk (vs. 21% for senior roles)
50% of junior design tasks could be eliminated (vs. 24% for art directors)
"Across the white-collar economy, entry-level jobs are suddenly vulnerable to automation because they involve low-stakes assignments of the sort that generative AI is best at." - Molly Kinder
And it's already happening. Revelio Labs has tracked a considerable decline in entry-level positions across industries. We're systematically eliminating the very roles that have traditionally taught young professionals how to work.
The Pattern
Consider how a young lawyer traditionally develops expertise. First year associates spend countless hours reviewing documents, drafting basic communications, and conducting research. It's tedious work, but it's how they learn to spot patterns, understand context, and develop judgment.
Now imagine all of that work handled by AI.
The same pattern is playing out everywhere:
Banks are debating cuts to incoming analyst classes
Marketing teams need fewer junior researchers
Even creative industries are questioning entry-level roles
Kinder's research at Brookings reveals consistent concerns across sectors: junior professors, freelance illustrators, novice auditors, and first-time TV writers all face the same threat. In interviews with dozens of executives, many admitted they're planning to hire fewer junior employees.
The Breaking Point
This isn't just about jobs disappearing. Robotics professor Matt Beane calls it the breakdown of the "expert-novice bond." Think about how expertise traditionally develops:
A young investment banker starts by building financial models. Tedious work, but it teaches them how companies actually function. They progress to helping with client presentations, learning how deals come together. Eventually, they're trusted with client interactions, having developed the judgment to handle complex negotiations.
AI is eliminating those crucial early and middle steps. It's like removing the foundation of a building while trying to add more floors on top.
The Perfect Storm
Here's where it gets interesting. At the exact moment we're breaking these learning paths, organizations are radically reinventing how they work.
Take Bayer, a 160-year-old life sciences giant with 100,000 employees. They're dismantling their entire hierarchy in favor of something called Dynamic Shared Ownership. No more traditional management. No more fixed roles. Instead, they're building:
Networks of autonomous teams that connect directly
Leadership roles focused on enabling rather than controlling
A "brand marketplace" where people choose projects every 90 days
The results are compelling. Their oncology team cut 800 unnecessary documents by asking one simple question: "What actually helps get this drug to patients faster?" Teams are spending 30% more time with customers. Bayer is targeting €2 billion in savings.
But here's the problem: How do you learn to be autonomous when the work that taught independence is automated?
The Warning Signs
When Matt Beane studied surgical training after the introduction of robotic surgery, he discovered something disturbing. Trainees spent more time watching and less time doing. The fundamental way expertise developed began to break down.
"If you watched enough hours of robotic surgery on YouTube," he asks, "did that make you ready to operate?"
Replace "surgery" with law, banking, design, or any other profession, and you see the pattern. We're creating organizations where:
AI handles the foundational work that once taught judgment
Hierarchies that guided development are vanishing
Traditional mentorship is being replaced by "watching from a distance"
Rishad Tobaccowala warns that "in less than 500 days the current waves of change will grow into a tsunami." But unlike most workplace transformations that primarily affect efficiency or productivity, this one strikes at something more fundamental: the development of human capability itself.
The Way Forward
As Tobaccowala points out, we're looking at 50-year careers in a world where companies last fewer than 15 years. The solution isn't to stop progress. It's to be intentional about what we're building.
Think of it this way: Instead of training musicians by having them play every note, we need to teach them to be conductors from day one. They need to understand the full score while directing AI to handle individual parts.
This means:
Reimagining how junior roles work
Turn AI supervision into a learning opportunity
Focus on developing judgment by overseeing automated work
Create apprenticeships in decision-making, not just task execution
Building learning networks, not laddersÂ
Like Bayer's "brand marketplace," create environments where:
Junior staff work directly with senior experts on real decisions
Learning happens through exposure to multiple challenges
Development comes through impact, not time served
Shifting our focus from skills to judgment
AI literacy becomes the new foundation
Communication becomes more crucial than technical skills
Creating and imagining trump executing and following
We won't know if we've failed at this for years - when we discover we've created organizations that can execute brilliantly today but can't develop the expertise they'll need tomorrow.
So before we rush to eliminate every entry-level task and dismantle every hierarchy, we need to answer one simple question:
What are we going to do about the young people?
I don't know, but I hope work will just change, rather than be removed, I believe similar things happened when the industrial revolution and computers came along, they did remove existing jobs, but new jobs came along.
Also VR or AR, could lead to apprentice style teaching for "doing" jobs like the surgeons etc