Few technological developments have provoked as much anxiety about the future of work as artificial intelligence. Opinion is sharply split between those who expect it to create more jobs than it destroys and those who foresee widespread unemployment. Having weighed both arguments, I believe the outcome hinges less on the technology itself than on how societies choose to manage it.
The optimistic camp draws reassurance from history. Every previous wave of automation, from mechanised looms to the personal computer, provoked similar fears yet ultimately spawned occupations that no one had anticipated. AI, its supporters argue, will likewise generate demand for specialists to design, train and supervise these systems, while liberating people from monotonous or dangerous labour and allowing them to concentrate on inherently human strengths such as creativity and empathy.
The pessimists counter that AI differs from earlier machines in a crucial respect: its reach extends into cognitive territory once considered uniquely human, placing not only manual workers but also professionals at risk. Their fear concerns pace rather than principle, namely that roles will vanish more quickly than displaced employees can plausibly retrain, and that the least educated will bear the heaviest burden, widening existing inequalities.
My own view is that these scenarios are not mutually exclusive; AI will simultaneously eliminate and create work, and the balance struck will be determined by policy rather than technology. Where governments and employers invest decisively in education and retraining, disruption can be absorbed. Left to market forces alone, however, the transition risks leaving a substantial minority stranded, which is precisely the outcome that deliberate preparation must avert.