Artificial intelligence (AI) is transforming manufacturing by improving efficiency, automating production processes and increasing competitiveness. In South Africa, where manufacturing contributes to economic growth, exports and employment for many semi-skilled workers, the rapid adoption of AI raises an important question: will it create new opportunities or place existing jobs at risk?

Although AI has been widely studied for its productivity benefits in advanced economies, far less is known about its effects on employment in developing countries such as South Africa, where structural unemployment, skills shortages and uneven technological readiness present particular challenges.

A recent South African study by researchers from the University of South Africa (UNISA) investigated how AI affects employment in the country’s manufacturing sector. The researchers examined whether AI adoption contributes to job displacement or supports employment growth over time. To achieve the research aims, the researchers analysed quarterly South African data from the first quarter of 2014 to the fourth quarter of 2024. The analysis included manufacturing employment, AI investment, education expenditure, manufacturing gross value added at basic prices, foreign direct investment (FDI) and wages.

Using a Structural Vector Autoregression model, the researchers examined both the immediate and longer-term effects of AI shocks on manufacturing employment. This allowed them to trace how employment changed over time following shifts in AI investment.

The study found that a positive AI shock resulted in an immediate and statistically significant decline in manufacturing employment. This negative effect persisted over the short to medium term, suggesting that AI adoption may initially displace workers in the manufacturing sector. By the tenth quarter, approximately two and a half years after the initial shock, AI explained nearly 30% of fluctuations in manufacturing employment. This indicates that AI becomes an increasingly important driver of employment changes over time.

The researchers also found that education expenditure, manufacturing growth and foreign direct investment were associated with higher manufacturing employment. In contrast, wage increases that were not matched by productivity gains were associated with lower employment.

Together, the findings suggest that South Africa’s structural unemployment, skills shortages and uneven technological readiness may make workers particularly vulnerable to AI-driven displacement unless complementary measures are introduced.

The researchers conclude that AI is becoming an increasingly important influence on manufacturing employment in South Africa. While it has the potential to improve industrial productivity, its adoption is associated with short- to medium-term job displacement in the manufacturing sector.

They argue that investment in reskilling, vocational training and AI technologies that complement rather than replace workers will be essential to support more inclusive and innovation-oriented industrial growth.