While much has been written about the incoming impact of AI and automation, limited work has been done on the potential impact of automation on economies in the developing world.
This prompted Daniel le Roux, a senior lecturer at the department of Information Science at Stellenbosch University, to investigate the situation in South Africa.
In an article for The Conversation, le Roux explained that he used data collected by Statistics South Africa for its Quarterly Labour Force Survey and an automation index produced by academics from the University of Oxford.
“From this, I was able to estimate that occupations performed by almost 35% of South African workers – roughly 4.5 million people – are potentially automatable in the near future,” he said.
“But the country appears ill-prepared for this reality. There is little discussion at policy level. Hardly any research has been done to investigate possible future scenarios.
“There’s also a great deal of uncertainty about how the uptake of automation technology may further drive inequality and preserve the asymmetry in the country’s economy,” he said.
According to le Roux’s data, roughly 14 million South Africans work in around 380 different occupation types.
64 of these occupations, employing an estimated 3.6 million workers, have a 90% or greater probability of being automatable in the near future, he said.
These occupations include, for example, cashiers, tellers, secretaries and telephone salesmen.
He found that the occupations of another 2.6 million workers, of whom 900,000 are employed as farmhands and labourers, have an 80%-89% probability of being automatable.
However he cautioned that workers of all skills levels are at risk, and that accountants, auditors and dental technicians aare extremely susceptible to automation.
“But trends suggest that people in low and medium-skilled occupations are generally more at risk than those who require extensive education,” he said.
“So it is not surprising that the country’s previously disadvantaged population groups are more exposed to job losses due to automation than their white counterparts.
“Half of all black South Africa workers are in occupations with an 80% or greater probability of automation; so are 47% of coloured workers. For white employees, however, the proportion is only around 30%.”
No preparations made
Despite the risk involved, le Roux said it was unlikely that South African business-owners would not take advantage of this new technology as it becomes available to them.
But there seems to be very limited high-level discourse about how South Africa plans to navigate this wave of technological advancement, he said.
“For South Africa, with its large number of low-skilled workers, a dramatically improved education system is an obvious and critical concern,” he said.
“Despite high unemployment, there remains a scarcity of skills in a wide variety of areas. This suggests a mismatch between supply and demand in the labour market.
“It is also important to understand how technologies will displace work in future. This understanding can help to better inform young South Africans’ career choices.”