There is a constant question about whether artificial intelligence poses a threat to jobs. Across boardrooms, policy circles and even dinner-table conversations, AI is often seen as a force that could take away employment, weaken middle-class work, and permanently change how people earn a living.
This fear now dominates much of the global debate around AI. But Ruchir Sharma tell us a different view. In his latest column for Financial Times, he says that the focus on job losses ignores an important reality at the same time, the world is running out of workers. Every major technological shift has created anxiety about the future of work.
From the steam engine to the internet, innovations have repeatedly raised fears that machines would replace people. However intense these fears were, they have never fully come true. What is often overlooked, Sharma says, is a powerful counterforce shaping today’s economy, a historic decline in the supply of labour.
The silent shock of shrinking populations
According to Sharma, over the past four decades, the number of countries with shrinking working-age populations has risen from zero to 55, now including most major economies. This change is no longer slow or distant. It is accelerating, as families are having fewer children than even pessimistic forecasts had predicted.
Recent data also shows the scale of the shift. China recorded its lowest number of births since the founding of the People’s Republic in 1949. Japan’s birth count fell to its lowest level since 1899.
Compared with projections made just ten years ago, the global fertility rate is now expected to fall below the level needed to stabilise population 25 years earlier than expected, around 2050. As a result, the world’s working-age population is projected to peak roughly 30 years sooner than once believed, around 2070.
AI and productivity
There are already signs that AI is raising output per worker, which on its own could reduce demand for human labour. But Sharma places this trend in the context of rapid population decline. In a world facing shrinking workforces, AI may be more likely to ease labour shortages than to cause mass unemployment.
History supports this view. Past technological revolutions did not eliminate jobs; they transformed them. Machines moved workers from farms to factories, and later from factories into services.
When cars replaced horses, carriage drivers did not disappear, they became truckers and cab drivers. In the 1970s, automated teller machines helped banks cut costs, open more branches and ultimately hire more human tellers. From the 1990s onward, one estimate suggests the internet displaced about 3.5 million US jobs but created nearly 19 million new ones.
Why this time it feels different
As per Sharma, Artificial intelligence may still prove more disruptive than earlier technologies. One forecast puts the maximum “addressable” labour market for AI and humanoid robots at four billion jobs worldwide. That sheer scale naturally fuels anxiety.
Job displacement has not occurred at a dramatic pace. Unemployment remains near multi-decade lows globally and in the United States, the world’s AI leader. In Japan, unemployment has continued to fall and labour force participation has risen, even as the use of robots has expanded rapidly.