The skill recession: How AI is quietly ‘untraining’ us

CULTURE & CODE

By Joey Briones

Let’s start with the obvious leadership paradox:

AI is making people work faster.
But, it is also making people less skilled.

And these two things are happening at the same time.

How do we know?

Because the World Economic Forum recently warned that 40 percent of workers’ core skills will be disrupted by 2030 due to artificial intelligence (AI) and digitalization.

And that disruption does not simply mean “new jobs, new skills.” It also means:

• the disappearance of entry-level roles
• the thinning of middle management
• the collapse of on-the-job learning
• the disappearance of mentors
• and the evaporation of the “ladder” new talent must climb (assuming the ladder still exists at all)

It’s ironic:

We spent decades trying to climb corporate ladders — and now AI is quietly removing the bottom rungs.

This is exactly what award-winning tech and AI researcherMatt Beane exposes in his book The Skill Code: How to Save Human Ability in an Age of Intelligent Machines (HarperCollins, 2024).

And with the rising trend of AI layoffs, stern warning also comes across – the more efficient companies become, the fewer opportunities humans have to actually learn.

It’s a strange moment in history.

We are working faster… while learning slower.

How Humans Actually Build Skill

For 160,000 years, human beings learned almost everything through one universal cycle:

Watch → Try → Fail Safely → Try Again → Earn Mastery

This ancient “expert–novice loop” was the foundation of every profession:

• surgeons learned beside seasoned surgeons
• chefs followed the head chef through chaos
• apprentices cut stone next to masters
• rookie engineers debugged systems with seniors
• analysts stayed late to learn the logic behind the numbers

Mastery wasn’t taught.

Mastery was absorbed — through proximity, practice, pressure, and feedback.

Then, AI arrived.

And with the subtlety of a runaway forklift, it inserted itself between expert and novice.

Generative AI’s Dirty Secret: It Makes Work Easier, but Growth Harder

Generative AI speeds up everything:

• report writing
• research
• code generation
• decision analysis
• customer responses
• content creation

Productivity shoots through the roof.

Executives applaud.

Shareholders beam.

Employees sigh with relief.

But here is the secret cost: When AI performs foundational work, novices lose the hands-on practice needed to build real skill.

You cannot climb a ladder if the first five rungs have been automated.

You cannot become an expert if the entry-level tasks no longer exist.

You cannot grow strategically if systems do all the cognitive heavy lifting for you.

This isn’t hypothetical — it’s happening now:

• Junior lawyers don’t review documents, because AI does it.
• Junior marketers don’t write drafts, because AI does it.
• Junior engineers don’t debug, because AI does it.
• Junior analysts don’t analyze — they “prompt” instead.
• Junior HR teams don’t screen — they read summaries.

So in “gym parlance” — generative AI is giving us speed, but stealing our “reps.”

And without reps, there is no muscle.
Without practice, there is no mastery.
Without challenge, there is no growth.

Meanwhile, Companies Are Cutting the Mentors Too

As if removing entry-level work weren’t enough, organizations are simultaneously:

• flattening hierarchies
• trimming middle management
• accelerating retirements
• cutting “cost centers” that include training and coaching
• replacing supervisory layers with AI workflows

The result?

The “expert–novice” relationship — the bedrock of human learning — is collapsing from both ends.

No entry-level tasks = no practice.
No middle managers = no coaching.
No senior talent = no exposure to judgment.

We are accidentally engineering “skill deserts.”

The Rise of the Shadow Learner (AKA: The Hungry Human Hacker)

Matt Beane discovered a fascinating phenomenon:

When organizations remove legitimate learning pathways, ambitious people simply create illegitimate ones.

Enter the Shadow Learner:

• the junior who sneaks into real work after hours
• the analyst who reverse-engineers AI output to see how it “thinks”
• the resident who positions themselves near the action to get a hand on the tool
• the engineer who secretly disassembles a robot to understand its soul

Shadow learners succeed despite the system.

But here’s the problem:
Shadow learning cannot scale.
Shadow learning is not a talent strategy.
Shadow learning is a cry for help.

And yet, it’s becoming the only route left to mastery.

When Efficiency Becomes a Threat to Capability

AI is incredible at boosting efficiency.

But efficiency is not the same as capability.

Efficiency removes friction.
Capability requires friction.

Efficiency cuts steps.
Capability requires steps.

Efficiency makes work smoother.
Capability requires rough edges.

Every leader must recognize the paradox:

If AI removes the struggle, it also removes skill development.
If AI removes the skill, it eventually removes the people.
If AI removes the people, the organization collapses in crisis.

Because here’s the part no AI evangelist likes to mention:

When something truly unexpected happens – AI freezes, but Humans solve.

Unless, humans no longer know how.

Leadership’s New Mandate: Protect the Skill Loop

If AI is becoming the operating system of work, leaders must become the designers of human learning.

It means intentionally re-creating the developmental rungs that AI has accidentally removed.

1. Restore Challenge

Give juniors tasks that stretch them — real tasks, not “AI babysitting.”

2. Rebuild Complexity

Let them see the whole workflow, not sanitized fragments.

3. Reintroduce Mentorship

Protect time for experts to coach novices.

4. Practice Judgment

Do not outsource all decision-making to AI.
Humans must still wrestle with ambiguity.

5. Bring Back Safe Failure

Create simulation spaces where mistakes are allowed —and expected.

If AI can’t fail in training, humans must.

6. Protect Entry-Level Work

Not everything should be automated.
Some tasks exist not for efficiency — but for development.

No rungs, no ladder.

The Organizations That Win Will Be Skill-Rich, Not Just AI-Rich

You can buy AI.
You can’t buy human capability.

AI improves automatically.
Humans improve only deliberately.

AI scales instantly.
Human skill scales only through culture.

AI learns from data.
Humans learn from challenge, complexity, and connection.

The companies that thrive will understand a simple truth:

AI is not your competitive advantage.
Your people’s ability to learn in an AI world is.

Machines are evolving.
The question is — are we letting humans evolve too?