Will Automation Destroy More Jobs Than It Creates?

Whether automation will ultimately destroy more jobs than it creates is a question that divides economists as much as the public, and both sides have historical evidence to draw on.

Those who fear net job losses point to the speed and scale of current automation. Unlike earlier technological shifts, which unfolded over generations, artificial intelligence is being deployed across entire industries within a few years, leaving little time for displaced workers to retrain. Roles in transportation, customer service and basic data processing are already being automated at scale, and critics argue that the new jobs created, largely in software engineering and AI oversight, require skills that most displaced workers simply do not have and cannot acquire quickly enough.

Optimists counter that this pattern has repeated throughout history without producing mass long-term unemployment. The Industrial Revolution destroyed countless manual jobs but eventually created far more in manufacturing, logistics and later services; the rise of computers eliminated typists and filing clerks but created entire industries around software and digital marketing. They argue that human needs are effectively limitless, so new industries will always emerge to employ people, even if we cannot yet predict what they will be.

In my view, both sides miss an important nuance: new jobs may indeed emerge, but the transition period will be painful and uneven, hitting certain regions and age groups far harder than others. A factory worker in his fifties is unlikely to retrain as an AI engineer, however many new jobs exist elsewhere.

Therefore, while I do not believe automation will cause permanent mass unemployment, governments must actively manage the transition through retraining programmes and targeted support, rather than assuming the market will resolve the disruption on its own.