Automation and the Changing Shape of Work

Few questions about AI attract as much attention as its effect on jobs. Two widely cited studies from international organisations offer a measured starting point — as long as their terms are read carefully.

Exposure is not replacement

Studies of AI and employment usually measure exposure: how many of the tasks in an occupation could, in principle, be performed or assisted by AI. A highly exposed job may be transformed, made more productive, or partly automated. Exposure alone does not say which of these will happen.

The ILO study

In 2023 the International Labour Organization published a global analysis of generative AI and jobs. It found that clerical work stands out: about 24 per cent of clerical tasks were assessed as highly exposed and a further 58 per cent as having medium-level exposure, which the authors describe as an upper-bound estimate. In other occupational groups the share of highly exposed tasks was much lower.

The study concluded that the main effect of the technology is more likely to be augmenting occupations than automating them, and that the larger impact may be on the quality of work — such as work intensity and autonomy — rather than on the number of jobs. Because clerical jobs are an important source of employment for women, the effects are expected to differ between women and men.

The IMF analysis

In January 2024 the International Monetary Fund estimated that almost 40 per cent of global employment is exposed to AI. The share is higher in advanced economies — about 60 per cent — than in emerging markets (about 40 per cent) and low-income countries (about 26 per cent). The IMF staff also estimated complementarity — how likely AI is to support rather than replace the people doing a job — and suggested that roughly half of the exposed jobs in advanced economies could benefit from AI through higher productivity, while the other half could see lower demand for labour.

What individuals can do

Studies of exposure point to a practical conclusion: tasks change faster than job titles. Understanding which parts of one’s own work are routine and which depend on judgement, relationships and responsibility is a useful way to think about where AI may help and where human skills will matter most.

Sources

1. Gmyrek, P., Berg, J., Bescond, D. Generative AI and Jobs: A global analysis of potential effects on job quantity and quality. ILO Working Paper 96, International Labour Organization, 2023.
2. International Labour Organization. Generative AI likely to augment rather than destroy jobs.
3. Cazzaniga, M. et al. Gen-AI: Artificial Intelligence and the Future of Work. IMF Staff Discussion Note SDN/2024/001, International Monetary Fund, 2024.