There’s a theory in modern economics that, despite the growing fears around AI’s encroachment on the labor market, technology creates more jobs than it wipes out. This may have been true for, say, the printing press or the internet. All the journalists, marketing specialists, and software engineers of the world are products of these technological developments. But this nugget of economic wisdom doesn’t survive the times. What is unique about artificial intelligence is not simply that it renders certain jobs obsolete — although that is rightfully a concern — it’s that it’s completely reorganized the ebbs and flows of the labor market.
Structural unemployment has traditionally been understood as a skill mismatch between workers and jobs. Get the right degree, hone the right skills, and you’ll be just fine. The invisible hand of the market is fair and rewards competence, after all. However, the onset of AI-driven hiring has fundamentally altered this equation, and with it, the nature of structural employment itself. The problem is no longer a skill mismatch because the gatekeepers aren’t really evaluating skills — at least not in any meaningful sense. The mismatch is between human capability and machine recognition, and the core skill being tested is algorithm appeasement.
With an estimated 90 percent of hiring teams using AI to sift through applications, the modern job search has become a master class in digital performance art. You sit hunched over your laptop at 2 AM, meticulously whittling down meaningful experiences and finely tuned, practiced skills into AI-friendly keywords. Every genuine accomplishment, every hard-won skill, every moment of growth gets compressed into sterile corporate-speak, all because before any human eyes see your application, it has to survive the gauntlet of the Applicant Tracking System.
This is a catastrophic failure of responsible innovation. When an
experienced project manager gets rejected because they wrote “
led cross-functional teams” instead of “managed stakeholder engagement,” that’s not a skills gap. When a talented designer’s application never reaches human eyes because their resume confused the parser, it’s structural exclusion by algorithmic design. Resumes were always imperfect proxies for capability, but at least when humans reviewed them, there was the possibility of seeing beyond the format and recognizing potential that didn’t fit the template.
If you’re lucky enough to be selected by the algorithm, you get an automated email (of course) back from the void inviting you to a dreaded HireVue interview. For those fortunate enough to be unfamiliar, HireVue is a platform that allows companies to conduct one-way video interviews. You sit alone in front of your webcam, reading questions off a screen, and record your answers while a timer counts down in the corner.
Video interviews have hollowed out interpersonal interactions into something clinical and deeply impersonal. The interview, traditionally a mutual evaluation where both parties assess fit and build a relationship, has been reduced to a one-way extraction of information. It’s effectively a data collection exercise. Using one-way interviews, companies don’t have to waste time with pleasantries or forge a relationship with a real person. To put it bluntly, they don’t have to know you. There’s an equation implied in the one-way interview paradigm: humanity subtracted, efficiency added, cost savings maximized
What’s particularly perverse about all this is that candidates are tested not on their ability to have an engaging conversation with a real person — arguably one of the most important skills in any job — but on their ability to talk at their own reflection with a ticking clock in the corner. It’s a fundamentally bizarre setup that bears little resemblance to actual work situations, and yet this is how we’ve decided to evaluate whether someone is qualified for a job that will involve countless human interactions.
The entire premise of division of labor is that specialization creates efficiency and allows people to pursue their particular talents and interests. The blacksmith becomes exceptional at smithing, the baker at baking, and society benefits from everyone doing what they do best. The market is supposed to be a system that recognizes and rewards specialized human capability. But when the sorting mechanism itself becomes fully automated, and when getting a job becomes less about demonstrating what you can uniquely contribute and more about appeasing the algorithm gods, that premise collapses entirely.
AI hiring systems, as currently deployed, automate the one part of hiring that most requires human judgment: the ability to see beyond programmed keywords and data, and to recognize unique value in individuals. The irony is that many of the innovators who built these very systems — with their unconventional backgrounds, non-linear career paths, and ability to see problems differently — would likely be filtered out by them.
Innovation is easy to celebrate. Even more so, it’s easy to dismiss its consequences as inevitable and necessary casualties. But innovation is a series of choices, and humans ought to be at the center of them.