A new Stanford study released last week, Canaries in the Coal Mine (Brynjolfsson, Chandar, Chen, 2025), analysed payroll data from ADP covering 25 million US workers across 7,000 job titles. They concluded from the data that since ChatGPT’s launch in November 2022, employment for 22–25 year-olds in the most AI-exposed jobs has fallen by 13%.
That’s the canary. Entry-level work built on codified knowledge is disappearing first.
The type of work once done by juniors: applying rules, following processes, filling templates. This is the very work AI now automates best.
But not all entry-level work looks the same. Some of it is rules-based and easy to replace. Some of it relies on judgment, context, and the ability to learn quickly in live situations.
This difference matters, because it explains why certain workers will stall while others accelerate.
From Canary to Divide
AI eats rules and processes. Anything you can find in a textbook, a manual, or a compliance checklist is now faster and cheaper in the hands of a machine. That is why new graduates, who walk into the workforce armed with codified knowledge, are the first to feel the squeeze.
Tacit knowledge is different. It is the judgment calls made in messy contexts, the workarounds discovered after years on the job, and the ability to read situations that do not fit the template. Older workers have this in their locker. AI does not (at least not yet).
This creates a widening gap. Workers who lean on routines and repeat the same task day after day will see their skills lose value and, over time, their employability. AI will do that work better, faster, and cheaper.
By contrast: the curious, those who tinker, push boundaries, and test what AI can deliver, will accelerate. They adapt faster, learn faster, and build tacit expertise sooner. For them, AI is not a threat but a multiplier. The same tools that erase rote work give them leverage to explore, experiment, and compound their skills.
One junior uses AI to churn out a passable draft in half the time. Another uses the same tools to test fifty variations, challenge assumptions, and uncover new options.
The first is standing still. The second is pulling ahead.
This is the curiosity divide: AI amplifies what is already there. The incurious become less employable. The curious become indispensable. And whether that divide strengthens your organisation or weakens it depends on how you design roles to reward curiosity.
The Mistakes Leaders Will Make
Leaders will trip up not because they misunderstand AI, but because they misread the roles around it (in-fact, maybe those are one and the same?). Too many will try to preserve junior jobs as they were, instead of redesigning them for the world we’re moving towards.
For the past century, the model was linear. Juniors carried the grunt work, seniors made the calls, and advancement came by mastering process step by step.
AI has broken that chain. Codified tasks won’t return, and nor should they. Paying humans for rule-based outputs that machines handle better is wasted effort.
That is why leaders who design tomorrow’s roles to mirror today’s are already behind. The jobs we build for the next decade will need redesigning again in another ten years.
They say the only constant is change. But that line doesn’t do justice to the pace we’re now living through. What once took a generation to shift now happens in twelve months.
So How Should Junior Roles Be Rebuilt?
The answer is not to protect junior jobs of the past, but to design new ones fit for an AI-driven future. The question shifts from “how do we save today’s jobs” to “how do we redesign junior roles for tomorrow’s workforce”.
In the AI era, juniors cannot be hired to push paper or churn boilerplate. Those tasks are already automated or augmented. What matters is curiosity, judgment, and building tacit skill from day one.
Think apprenticeships, not grunt work. Juniors as system probers, workflow testers, and context gatherers. Their job is not to be the knowledge base, but to interrogate it, push its boundaries, and turn discovery into insight through practising curisoity.
The beauty is that we all start with this skillset. As kids, curiosity is natural. But our education system arguably conditions us to trade discovery for correctness, rewarding right answers and punishing mistakes. Over time, curiosity fades, and ego hardens around the need to be right.
Individuals have to keep practising curiosity, but they cannot do it alone. Workplaces either reinforce it or erode it. Organisations that design roles around compliance will watch juniors stagnate. Those that design for curiosity will see juniors compound tacit knowledge and move up the ladder with speed.
Because the centre of gravity has already shifted. AI handles the knowledge work we once got rewarded for. What remains is wisdom work. Wisdom work is what curiosity builds toward. It is the power skillset AI cannot replicate:
Knowledge work = writing a perfect strategy doc. Wisdom work = spotting where your team will resist and defusing it upfront.
Knowledge work = being right in a meeting. Wisdom work = creating the space where the best idea can win, even if it is not yours.
Knowledge work = summarising 40 pages. Wisdom work = naming the insight no one wants to say out loud.
These are not soft skills. They are power skills. Curiosity is the entry point, because in the wisdom era, value comes from knowing which questions matter, when to act, and how to move people forward without needing to control them. That is the trajectory junior roles should be built for.
Three Redesign Principles
These principles are not just about reworking job descriptions. They are about closing the curiosity divide inside your own business and building pathways into wisdom work.
1. Apprenticeship over grunt work
Pair every junior with a named mentor.
Make tacit transfer explicit: shadowing, post-mortems, “show the edge cases” sessions.
Shift performance reviews away from output volume and toward rate of skill acquisition: how quickly juniors can apply judgment in new contexts.
2. Exposure over siloing
Stop burying juniors in back-office tasks for two years.
Give them early visibility into client conversations, decision trade-offs, and the “why” behind strategy.
A simple rule: within their first six months, every junior should present back to the team what they’ve learned from observing live decisions.
3. Curiosity over compliance
Don’t reward juniors for rule-following that AI already automates.
Set challenges that require them to test AI, not just use it — probe its limits, find errors, suggest workflow tweaks.
Track curiosity explicitly: are they raising better questions, proposing improvements, stress-testing the system?
Done right, this won’t preserve junior jobs as they were. It will transform them into accelerators of tacit knowledge: capability machines cannot replicate.
Ignore this, and the early cost savings will vanish. Juniors will disappear, mid-level managers will dry up, seniors will retire, and what remains will be a hollow firm: efficient on paper, fragile in practice.
The Closing Signal
The Canaries in the Coal Mine study is more than an early warning. It is the first hard evidence that AI is reshaping the workforce from the bottom up. The 13% drop in junior employment is the clearest sign yet: codified, rule-based entry roles are gone for good.
And here is the blunt truth: the biggest barrier to getting or keeping a job will not be AI itself. It will be a lack of curiosity. Those who keep leaning on routines will drift into irrelevance. Those who use AI as a tool to test, probe, and discover will accelerate. The incurious will find fewer doors open. The curious will find more opportunities than ever.
AI has not ended entry-level work. It has ended entry-level work as we knew it.
The firms that redesign around curiosity, judgment, and tacit learning will not only rebuild the ladder stronger, they will close the curiosity divide. The ones that fail will find the ladder gone, along with their future talent pipeline.
For firms, curiosity must be the organising principle of junior roles. For workers, curiosity is the new entry requirement.
Written by Mike ✌
Passionate about all things AI, emerging tech and start-ups, Mike is the Founder of The AI Corner.
Subscribe to The AI Corner
