By Veronika Bak, PhD Candidate, Department of Media and Communication, Erasmus School of History, Culture and Communication, Erasmus University Rotterdam, the Netherlands
This blog is produced as a part of the Reshaping Work Fellowship Programme. The opinions and views expressed in this publication are those of the author. They do not purport to reflect the opinions or views of Reshaping Work or organisations that have supported the programme.
Think about the last time you watched someone learn a skill by doing it — not from a screen, not from a manual, but shoulder-to-shoulder with someone who already knew how. Now ask yourself: when did you last see that happen in a workplace that has adopted AI?
We are in an era where virtually every workplace acknowledges that new technologies demand new skills. But the dominant response has been to layer formal upskilling programmes on top of work, rather than to ask whether the work itself still teaches. Those closest to the work have the knowledge to understand what is really happening, but organisations are rarely structured in ways that allow this knowledge to surface. Dr. Matt Beane’s The Skill Code: How to Save Human Ability in an Age of Intelligent Machines provides compelling evidence for exactly this point, reframing it as a systemic problem rather than an individual one.
The debate about AI and work has been dominated by job displacement: how many roles will disappear, which tasks will be automated, who will be left behind. Beane reframes the problem entirely. Across hospitals, factories, law firms, and military units, the technologies we adopt in the name of productivity are quietly severing the link between experienced workers and the people who are supposed to learn from them. Beane names this pattern and, more importantly, offers a way forward.
Dr. Beane, assistant professor at UC Santa Barbara and a Digital Fellow at both Stanford and MIT, spent over a decade studying workplaces where intelligent technologies are already deeply embedded — surgical suites, warehouses, chip-design labs, military bomb disposal units, law firms, and more. His core finding: across all of these contexts, the way we are deploying technology is quietly destroying the conditions under which people develop real, reliable expertise.
The three Cs framework and what it draws on
His argument rests on three building blocks he calls the three Cs: Challenge, Complexity, and Connection. Challenge refers to working near the edge of one’s cabability — struggling enough to learn, but not so much to fail entirely. Complexity refers to engaging with the broader context of the work, not just a narrow task, so that understanding grows alongside skill. Connection refers to the bonds of trust and respect between expert and novice that make access to challenging, complex work possible in the first place.
The three Cs provide a synthesis: a way of naming what decades of research in psychology, education and organisational sociology have established about how people actually build expertise. Beane adds a systemic account of how technology disrupts all three conditions simultaneously across a wide range of very different workplaces. This framework aims to give organisations a concrete vocabulary for diagnosing what is going wrong.
Stories that stick
Beane’s great gift as a writer is the vividness of his case studies. Kristen, the surgical resident who stands aside while a robot-assisted procedure unfolds without her; Sita, the warehouse supervisor who understood the entire system by observing what others had overlooked and pursuing the questions her supervisor had suggested; and Inés, the factory worker who cracked the error-code logic of a robotic line and proposed a solution that was implemented without her receiving any credit or recognition. These are not abstractions. They are people whose working lives are shaped, and sometimes diminished, by technological decisions made far above their level.
Shadow learning
The chapter on shadow learning is the one most likely to change how readers think about their own organisations. Beane finds that only a small minority of workers in highly automated environments, perhaps one in eight depending on the context, find workarounds to obtain skills the organisation no longer teaches them: watching YouTube videos after hours, spending time with a colleague from another department, lingering near a machine after clocking off.
What makes this concept so striking and, for a future-of-work audience so urgent, is what it reveals about the gap between how organisations think learning happens and how it actually does. Shadow learning is the signal that the work itself has stopped teaching, and most workers have neither the time nor the inclination to find a way around it.
Consider Inés, who understood the robotic system better than the engineers who designed it, but whose insight was captured only by accident. Shadow learners like her represent exactly the kind of frontline knowledge that managers miss when evaluating whether a technology implementation is working. If learning has gone underground, the problem is not the learners, it is the design of the organisation they work in.
Why this matters for HR and policy?
For HR leaders and policymakers, the The Skill Code raises some uncomfortable questions. If the conditions for skill development are being quietly eroded by the very productivity tools organisations are investing in, then talent pipelines are more fragile than they appear. The senior experts who drive results today were shaped by a different, richer learning environment. When they retire or move on, will there be anyone ready to replace them?
Beane’s answer is that this is a choice, not a fate. The technology does not determine the outcome, the human decisions about how to deploy it do.
The bottom line
The Skill Code deserves a place alongside the most serious recent contributions to the AI-and-work conversation, For researchers, HR professionals, policymakers, and organisational leaders, Beane offers both a diagnosis and a direction. The ‘skill code’ is under threat. If we choose to act on it, it can be rebuilt.
The Skill Code: How to Save Human Ability in an Age of Intelligent Machines is available from Harper Business (2024). Matt Beane’s research can be found at mattbeane.com.