EMILE ORMOND: SA is not ready for AI-era job disruption

As artificial intelligence (AI) grows more sophisticated and pervasive, its potential to disrupt labour markets demands urgent attention. There have been numerous predictions and expert claims that AI will be highly disruptive to the workforce.

For instance, McKinsey estimates 400-800-million people globally may need new jobs by 2030, while a BCG survey found 42% of workers fear their roles may vanish within a decade. 

For SA, with an unemployment rate of 32.9% and 46.5% for youths, these predictions are dire. The country simply cannot afford large-scale job losses without jeopardising fragile social stability and deepening poverty and inequality. As the government of national unity prioritises “inclusive growth and job creation”, understanding AI’s impact on jobs is not just critical — it’s urgent. 

Impact yet to materialise 

Evidence of current AI-driven job losses remains limited. In advanced economies such as the US and EU, unemployment is near historic lows. Research has found that for now AI’s impact on employment is minimal, often boosting productivity instead. So far, AI has not significantly shrunk job markets globally or locally. Historical technological leaps, like the industrial revolutions, sparked similar fears of mass labour market disruption but ultimately resulted in substantially higher employment and productivity. 

This precedent, combined with AI’s limited impact to date, may have bred complacency among South Africans that AI’s impact will be manageable and a net positive. However, this view is shortsighted. Rapidly increasing advances in areas such as multi-modal and agentic AI are poised to transform workplaces. The vast majority of organisations are planning on introducing or expanding their use of AI. This will see workers requiring new skills, creating new roles and eliminating others. While the balance of these changes is debated, huge labour market disruption is almost certain. 

This time is different 

AI’s unique traits, distinct from past technologies, will amplify its impact on jobs. These features include: 

• Cognitive capabilities. Unlike earlier automation that targeted manual tasks, AI can handle complex cognitive work. 
• General-purpose technology. Like electricity, AI’s application spans all sectors, driving broad economic impact and broadly fuelling productivity at an unrivalled pace. 
• Self-improvement. AI can help enhance future iterations of itself, unlike previous technologies. Essentially, AI can make AI better. 
• Democratised access. Many AI tools are freely or cheaply available. 
• Rapid adoption. For example, generative AI surged from obscurity to global prominence in just three years. 

These characteristics illustrate why AI will disrupt labour markets at an unprecedented pace and scale, but not all countries and groups are equally vulnerable. 

SA has some breathing room 

High-income countries, with more white-collar jobs, face earlier AI-driven disruption. For instance, 34% of EU jobs are exposed to AI automation, compared to 19% in the AU, according to the International Labour Organisation (ILO). Aging populations and high labour costs may also accelerate AI adoption in developed markets.

Young workers, often in entry-level roles, are particularly at risk, potentially blocking their entry into the labour force. In other words, emerging economies such as SA may be partially shielded from the initial waves of AI automation, but when it inevitably arrives the country could be especially vulnerable due to its large young labour force.

In conjunction with this, AI will likely also drive significant productivity gains and create new, currently unforeseen jobs, but the transition period could be long and hard and ultimately further entrench SA’s world leading inequality. 

Charting a path forward 

SA has a narrow window, as short as two to three years, to harness AI’s productivity gains while mitigating its fallout. Key actions stakeholders can take include: 

• Policy development. Political leaders must move beyond vague rhetoric and adopt nuanced, thoughtful policy positions on AI, such as concluding the long-awaited national AI strategy. 
• Digital infrastructure. Expand reliable, high-speed internet nationwide, resolving disputes over providers like Starlink to ensure equitable AI access. 
• Reskilling programmes. Invest in large-scale training to equip workers with AI-relevant skills and update school and tertiary education curricula for emerging roles.
• Responsible AI governance. Regulators and organisations should integrate AI oversight into corporate governance, aligning innovation with national development goals. Moreover, AI needs to be a cross-cutting responsibility in government. 
• Social protections. Plans for displaced workers need to be considered now — there are nearly 19-million grant recipients, compared to a tax base of 7-million. Growth measures and/or new revenue sources will need to be found if the 38% of tax revenue coming from personal income tax is in jeopardy. 

SA stands at the edge of an epoch-defining labour shift. The question is whether we act proactively or react in crisis. 

• Ormond has a PhD in AI ethics risk governance and two decades of experience in policy analysis.