AI can now do or accelerate many of the tasks that once filled the early years of a career. That is good for productivity. But those repetitive tasks were also where people learned to recognize patterns, catch mistakes and develop the judgment we later expect from senior employees.
By Sherri Claydon | WBN News - Vancouver | September 18, 2026
Editor: Karalee Greer Subscription to WBN and being a Contributor is Free
AI can now do or accelerate many of the tasks that once filled the early years of a career. That is good for productivity. But those repetitive tasks were also where people learned to recognize patterns, catch mistakes and develop the judgment we later expect from senior employees.
The question is not whether we should preserve inefficient work. It is what happens to the learning when the work disappears.
There is something experienced people do that can be difficult to explain.
They look at a piece of work and know something is wrong.
Sometimes they cannot immediately tell you why. They have seen enough versions of the same problem, made enough mistakes, corrected enough errors and encountered enough exceptions that something simply does not fit.
We call that judgment, experience or instinct. But it did not appear when someone became senior. It was built over years of doing the work.
Junior developers wrote code and learned what broke. Junior lawyers researched cases, reviewed documents and produced first drafts. Junior accountants worked through audit files, confirmations and transactions.
A lot of that work was repetitive. Some of it probably should disappear. But repetition was also how people encountered patterns and exceptions often enough to recognize them later.
We still want the judgment. We just may be removing some of the experiences that traditionally built it while asking people to demonstrate that judgment earlier.
PwC calls the shift the “seniorisation” of entry-level work.
Its 2026 Global AI Jobs Barometer analyzed more than one billion job advertisements across six continents. An analysis of 2.4 million U.S. entry-level jobs found that roles most exposed to AI were seven times more likely to require traditionally senior-level, human-intensive skills such as leadership, creativity and face-to-face interaction.
Job openings for these “seniorised” entry-level roles grew 35% from 2019, while other entry-level roles declined 10%. PwC says skills such as judgment, creativity and leadership are becoming more important as AI takes on routine work.
That creates an interesting contradiction: we may be removing some of the work through which people traditionally developed judgment while simultaneously asking them to demonstrate that judgment sooner.
This is a different question from whether AI will eliminate entry-level jobs. It is about what happens to the career path when some of the work at the beginning changes.
This is not a distant issue for British Columbia.
Statistics Canada found that 25% of workers in B.C. reported using generative AI at work during the previous year, based on surveys conducted from September 2024 through July 2025. That was the highest reported proportion among the provinces and regions studied, ahead of Ontario at 24%, Quebec and Alberta at 21%, Atlantic Canada at 18%, and Manitoba and Saskatchewan at 17%.
For Vancouver businesses, that makes the question particularly relevant. If AI adoption is already changing how work gets done, employers also need to think about how the next generation of experienced professionals will develop the judgment their businesses will eventually rely on.
Globally, the World Economic Forum says more than one in three young workers are employed in occupations with medium-to-high exposure to AI-driven task change. Its June report, developed with PwC, argues that organizations need to rethink job design and talent pipelines as AI changes early-career work.
Software development offers an interesting example because the evidence does not fit neatly into a story about AI hurting junior workers.
Researchers combined randomized field experiments at Microsoft, Accenture and an anonymous Fortune 100 company involving 4,867 software developers. Developers given access to an AI coding assistant completed about 26% more tasks, and less-experienced developers had both higher adoption rates and greater productivity gains.
That matters because AI can clearly help less-experienced employees become more productive. What we know much less about is whether working that way also helps them build the knowledge and judgment they will need to operate independently at a senior level.
The legal profession is starting to ask that question more explicitly.
The American Bar Association recently called it “The Apprenticeship Problem in the Age of AI.” Legal research, document review, drafting and preparing case chronologies are among the tasks AI can increasingly perform or accelerate. They are also tasks that historically helped junior lawyers build skills and expertise.
Paul Saunders, chief strategy and innovation officer and partner at Halifax law firm Stewart McKelvey, calls it the “AI training conundrum.” He points to work such as reviewing leases and researching cases that once required junior lawyers to work through the material themselves. AI can now summarize hundreds of leases or help produce a research memo, leaving the junior lawyer to validate the output.
The efficiency gain is easy to see. The harder question is how someone learns to recognize a bad summary if they have not had enough opportunities to understand what good work looks like.
The answer may not be to preserve the old way of learning.
A major study by researchers Erik Brynjolfsson, Danielle Li and Lindsey Raymond examined the rollout of a generative AI assistant to 5,179 customer-support agents.
AI increased productivity by 14% on average, with gains of 34% for novice and lower-skilled workers. The researchers also found suggestive evidence that AI was spreading practices associated with more capable workers and helping newer employees move along the experience curve.
Maybe AI does not simply remove learning opportunities. Maybe it can transmit expertise too.
We do not yet know. Stanford Digital Economy Lab is currently studying that question through a randomized controlled trial with final-year IT apprentices in Germany. Researchers are testing whether AI enables apprentices to perform more advanced tasks and whether that improved performance comes at the expense of comprehension. The project page does not yet report results.
The distinction between doing something and understanding it matters. But it does not mean we need to preserve repetitive work just to preserve the learning.
If reviewing 100 documents taught someone to recognize patterns and exceptions, perhaps they do not need to manually review the next 100. They need enough exposure to know when document 101 does not fit the pattern.
If writing basic code taught a developer how different parts of a system connect and where things tend to break, AI may be able to write some of that code. The developer still needs to understand the system well enough to recognize when the output is wrong.
The repetitive task may deserve to disappear. The learning outcome does not.
Once we separate the two, a more interesting question emerges: could we use AI to help people develop that judgment better, and perhaps faster, than they did before?
There are already examples of what that could look like.
Deloitte is redesigning its graduate audit program in England and Wales as AI changes the manual work traditionally performed by junior accountants. Beginning this month, trainees will complete most of their professional exams in the first year while using simulations that combine traditional and AI-assisted auditing. The goal is not to preserve the old tasks, but to prepare people for judgment-based work earlier.
Deloitte Canada takes the idea further, arguing that learning through work needs to become more intentional as AI and hybrid work reduce some of the repetition and proximity employees once learned from. Experienced employees can make their reasoning visible, create opportunities to practise and give feedback in real time. AI can help create scenarios and questions that test an employee's thinking.
That changes the questions businesses need to ask.
When AI can take over a junior task, “Can we automate this?” is only the first question. The deeper one may be: What did someone learn by doing this task, and how will they learn it now?
The impact of AI on work will not be measured only by how much time it saves. For Vancouver businesses already adopting these tools, there is another question: who is developing the judgment they will need five or ten years from now?
Businesses need people who can recognize when an answer is wrong, understand why it is wrong and make decisions when there is no obvious answer at all.
AI may actually give us a better way to build those capabilities. Instead of relying on years of repetition and hoping judgment emerges, businesses can expose people to meaningful problems earlier, create deliberate opportunities to practise and make the thinking of experienced employees more visible.
That changes more than training. It changes how we think about experience itself.
Perhaps the future career ladder is not one where people spend years doing simpler work before they are trusted with harder problems. It may be one where AI handles more of the repetition while people encounter complexity sooner, with the support to learn from it.
The opportunity is not to recreate the way we learned before AI. It is to build a better way to learn because of it.
By Sherri Claydon | WBN News - Vancouver LinkedIn: https://www.linkedin.com/in/sherriclaydon
Editor: Karalee Greer Subscription to WBN and being a Contributor is Free
Source Information
Tags: #Sherri Claydon #WBN News #Artificial Intelligence #Future Of Work #Entry Level Work #Workplace Learning #Skills Development #Professional Development
