The Hidden AI Risk for HR: The Collapse of the Career Ladder

HR leaders: I think we may be asking the wrong AI question.

The question is not just:

“Which jobs will AI replace?”

It may be:

“What happens when AI removes the work that used to develop people?”

Because that is where the research gets interesting.

Across the major studies, the pattern is fairly consistent:

• AI is mostly reshaping tasks, not simply wiping out whole jobs

• around 20–30% of tasks in many roles are likely to be automated, while the rest are augmented

• the biggest disruption is skills mismatch

• and entry-level / routine knowledge work is among the most exposed.

That last point should matter a lot to HR.

Because many of the tasks AI is absorbing are exactly the ones junior employees used to learn from: basic analysis, drafting, structured admin, document review, process-heavy support work.

Not exciting work...

But often the work that built:

• judgment

• business context

• confidence

• and readiness for bigger roles.

So the real risk may not be mass job loss. It may be the quiet hollowing out of the career ladder. And if that is true, then the HR challenge changes completely.

It becomes less about headcount planning alone, and more about:

• how we redesign junior roles

• how we replace lost apprenticeship experiences

• how we protect critical thinking

• how we support managers now expected to lead both people and AI-enabled work

• and how we stop “AI efficiency” becoming a synonym for “less development”.

The research also points to some other issues HR cannot ignore:

• a growing trust gap around AI systems and how decisions are made

• the risk of cognitive atrophy if people stop challenging outputs

• and productivity debt, where always-on AI creates pressure for humans to work at machine pace.

So here is the provocation:

If AI is taking away the bottom rungs of the ladder, what is your organisation doing to build a new one?

Curious whether others in HR / Talent are seeing this yet.

Are you already redesigning early-career pathways and manager support around AI — or does this still feel like a future problem?