Technological Unemployment

Technological unemployment is the loss of employment caused when new technology replaces human labour in production. AI-driven automation creates it when machines take over enough workplace tasks that displaced workers cannot be absorbed promptly into new or expanding jobs. It may be temporary during adjustment or persistent where job creation and reskilling remain inadequate.

How displacement occurs

AI can perform cognitive tasks such as prediction, classification, content generation and decision support. When firms reorganise production around these capabilities, automation can affect both manual and white-collar work.

• A firm substitutes AI-enabled capital for workers when automation performs their tasks more cheaply, rapidly or consistently.
• Jobs containing many routine and codifiable tasks are generally more exposed because their procedures can be represented through data and rules.
• AI may automate only part of an occupation, but the resulting productivity increase can allow the same output to be produced with fewer workers.
• Automation can also reduce employment indirectly in suppliers or occupations whose services become less necessary.

When displacement becomes unemployment

Job displacement becomes technological unemployment when workers released by automation cannot move quickly into expanding activities. Thus, the problem depends not only on jobs lost, but also on the pace and inclusiveness of labour reallocation.

• A skill mismatch arises when new jobs require capabilities that displaced workers do not possess.
• Sectoral and geographical mismatches can prevent workers from accessing newly created employment.
• Rapid adoption can make retraining and job creation slower than displacement, producing transitional unemployment and downward pressure on wages.
• Unequal access to digital infrastructure, education and capital can concentrate productivity gains while widening labour-market inequality.

Why the outcome is not predetermined

AI can complement workers as well as replace them. Higher productivity can lower costs, raise demand and create new tasks, occupations and industries, so net employment depends on whether these effects outweigh displacement.

• Human-centred adoption should use AI as an augmenting technology, especially where judgement, accountability and interpersonal skills remain important.
• Education, continuous skilling, worker mobility and social protection can shorten unemployment spells.
• Domestic innovation and diffusion of affordable technology can spread productivity gains, while support for employment-intensive sectors can expand alternative work opportunities.