Generated by Codex with GPT 6 Astra High

Artificial intelligence can replace tasks while increasing demand for workers elsewhere. In “Help wanted”, The Economist argues that this second effect is already substantial: its estimate puts American jobs created by AI at roughly one million, compared with about 200,000 announced AI-related layoffs since mid-2023. The article describes an uneven employment expansion, with losses in routine office work alongside demand for people who build, deploy and use the technology.

Three routes to more work

The most tangible gains come from infrastructure. Data centres need electricians, cooling specialists, equipment makers and engineers, as well as computing hardware. The magazine finds roughly 320,000 additional jobs since 2023 across five related industries, measured against broader construction and manufacturing trends. It acknowledges that other factory projects and electricity-grid upgrades account for some of the increase. Even so, the competition for workers is visible: Indeed reports that data-centre installation and maintenance vacancies advertise pay around 40% above comparable work elsewhere.

A second source is the work of making AI useful. Companies need model builders, people who evaluate training data and outputs, engineers who adapt systems for customers, and managers responsible for adoption. LinkedIn estimates that about 640,000 AI-specific jobs were added between 2023 and 2025. Separately, The Economist finds employment in selected technical professions roughly 730,000 above the growth implied by professional employment overall. These are different, potentially overlapping measures; they should not be added together as independent job totals.

The third mechanism is less direct. When AI makes a service cheaper to provide, customers may buy enough more of it to increase total employment. The article points to paralegals and market-research analysts, whose employment rose by about 11% and 6% respectively between 2023 and 2025, against roughly 2% nationally. Those increases are consistent with productivity expanding demand, though they do not by themselves prove that AI caused the growth.

Aggregate gains can hide individual losses

Routine administrative work faces a harsher adjustment. The article reports that, since January 2023, customer-service employment has fallen by about 10%, while employment among secretaries and administrative assistants has dropped by roughly 15%. Growing demand for engineers or electricians does not automatically provide a route into those jobs for displaced office workers.

The central takeaway is that counting announced layoffs captures only one side of technological change. Investment, new occupations and expanding demand can offset automation’s losses. The evidence presented supports that possibility today, while leaving open how the balance will change as AI becomes more capable and the infrastructure boom matures.

Source: The Economist, September 12th 2026, “Help wanted”, Finance & economics, pp. 65-66; listed in the contents as “The AI jobs boom”.