Hrpro.be
Much of the public debate about AI and work focuses on a blunt question: which jobs will disappear?
A more interesting—and more uncomfortable—question is what happens after a job is disrupted. Who actually manages to land on their feet, and why?
Two recent academic studies offer a useful lens on this. Together, they suggest that labour-market transitions are shaped not only by economics or skills, but also by psychology, identity, and unequal starting positions.
The first study, by Sam Manning and Tomás Aguirre from GovAI, looks at how exposed different occupations in the US are to AI—and, crucially, how capable the people in those jobs are of adapting if disruption occurs. The researchers assessed 356 occupations along two dimensions: how easily AI could perform the tasks involved, and how “adaptable” the workers in those roles are. Adaptability here is not a vague personality trait, but a composite of concrete factors such as age, geographic mobility, financial buffers, and how transferable someone’s skills are across sectors.
One striking finding is that AI exposure and adaptability often move together. Many of the jobs most affected by AI—such as digital design, data analysis, or finance—are held by workers who are relatively well positioned to cope with change. Out of roughly 37 million Americans working in the most AI-exposed occupations, about 26.5 million are in roles where people typically have the resources, skills, and flexibility to pivot if needed.
But that still leaves a substantial group behind. More than six million workers fall into a much more precarious category: they are highly exposed to AI and have low adaptive capacity. These are often clerical, administrative, or routine coordination roles—jobs that are vulnerable to automation but do not easily open doors to adjacent careers. The study is careful about its limits: AI exposure does not automatically mean job loss, and these are aggregate patterns rather than individual destinies. Still, the message is hard to ignore. Technological disruption does not hit a level playing field.
If the first paper highlights structural inequality, the second zooms in on something more subtle: how people feel about changing careers.
A study by Alexia Delfino and colleagues at Bocconi University focuses on reskilling decisions among unemployed jobseekers in Italy. The researchers surveyed over 1,100 people, asking them to choose between two hypothetical training options. One was a generic course that improved employability without implying a career switch. The other was a reskilling programme designed to move participants into a new profession.
The twist was that the reskilling option randomly assigned people to one of two in-demand roles: IT assistant or construction technician. Both were framed as white-collar, office-based jobs, and both offered better long-term prospects than the generic course.
From a purely economic perspective, reskilling should have been the obvious choice. Yet participants split almost evenly between the generic and reskilling options. What tipped the balance wasn’t income expectations, but identity. People were significantly more likely to choose reskilling when the target role was in IT rather than construction. The decisive factor was whether the new job “felt right”: whether it matched how people saw themselves and how they believed the job was viewed socially.
In other words, labour-market transitions are not just technical problems of skills mismatch. They are also deeply psychological. People ask themselves not only “Can I do this?” but “Is this who I am?”
Taken together, these studies point to a sobering conclusion. If AI accelerates job transitions—as many expect—it will not be enough to offer training programmes, incentives, or optimistic narratives about lifelong learning. Some workers start with more room to manoeuvre than others. And even when opportunities exist, people may resist them if they threaten their sense of identity or status.
The future of work, then, is not only about technology or economics. It is about adaptability, inequality, and the quiet, human question of whether a new path feels like a step forward—or a step down.
These two studies were referred to in The Economist - Bartleby.
