Rob RoseThere was an audible gasp in the audience at the Inanda Club in Sandton when Larry Fink, the CEO of BlackRock, described the impact that AI has had on jobs at his firm, the largest asset manager in the world.
“In the past five years, we grew in total about $4-trillion, and we haven’t really added any new employees,” said Fink, a shift he attributed entirely to adopting AI and data in its business.
Adrian Gore, the CEO of insurance company Discovery, who was interviewing Fink at the event last week, raised his eyebrows, evidently factoring in the implications of this statement for a jobs-scarce country like South Africa.
“I do believe AI will become the next Industrial Revolution,” Fink continued. “We are using [it] for productivity gains and, more importantly, to get better insights to invest better on behalf of our clients.”
It’s not just BlackRock, a company with $10-trillion in assets under management. In August, Doug McMillon, CEO of the world’s largest retailer, Walmart, gushed about how using generative AI had been instrumental in improving more than 850-million pieces of data in its product catalogue.
“Without the use of generative AI, this work would have required nearly 100 times the current headcount to complete in the same amount of time,” McMillon said in Walmart’s earnings report.
For a country like Japan, facing the prospect of its population shrinking 16.3% until 2050 and with unemployment of just 2.4%, or Italy (population set to decline 10.1% by 2050), these productivity gains sound like a marvellous way to keep GDP growth ticking along.
But for South Africa – a country with a morale-sapping unemployment of 42.6% if you include those who have given up looking for work and a population expected to grow more than 10% to 68-million by 2050 – the prospect of an AI-led jobs purge seems positively dystopian.
Consulting firm Accenture predicted in 2018 that, in South Africa, 5.7-million jobs are “at risk” due to automation – 35% of all jobs. “South Africa is less prepared than other countries and needs to give its workforce skills to participate in the digital economy,” said director Roze Phillipps at the time.
If anything, the trajectory has accelerated. The World Economic Forum wrote in its Future of Jobs report last year that thanks to various factors, including technology, of 673-million jobs it surveyed, it expected a net decrease of 14-million jobs – with 83-million jobs cut, and 69-million new roles created.
Those who responded to its survey predicted that “42% of business tasks will be automated by 2027”.
A strategy based on simply hoping to somehow remain standing after this global tsunami hits, in other words, is no strategy.
An ethical question
Discovery Bank chair Reuel Khoza said Fink’s view on growing companies without creating jobs raises serious ethical questions.
“What is the purpose of growth, and growth that does not provide employment and livelihoods to ever-increasing populations?” he asked. “AI does help a great deal – it is at the cutting edge of growth and technology – but it raises human and humane questions which continue to be vexing.”
Others were less concerned.
Eskom chair Mteto Nyati argued that this ought to be reframed as an opportunity: South Africa’s unemployment rate among the youth is north of 60%, yet this is the demographic that not only has greater critical thinking skills but is also best placed to learn digital skills.
“For our companies to grow, the thing that will limit that growth is the fact that we don’t have digital skills [while] we’ve got people trained to think sitting idle,” he said.
The problem is that South Africa doesn’t have any choice but to embrace AI.
Hendrik du Toit, the CEO of Ninety One, who also watched Fink’s presentation, said any country that spurns this new technology risks finding itself out in the cold, unable to integrate into the global economy.
“We are at a stage in the world similar to the Industrial Revolution where a little island called the UK became disproportionately powerful because of one invention, the steam engine,” he said. “If I have a fear this morning, it’s that we talk about ourselves and forget the world is moving on really fast.”
The US, he said, is now building a moat around itself of technological capacity and an ability to process immense amounts of data; countries that don’t stay in step with new technology are making themselves irrelevant.
“We’ve got to get on this bus and understand what the opportunity is and [what the] risk is,” said Du Toit. “We need to find a way to stay integrated.”
Raising inequality
So, how prepared is South Africa for this tsunami?
Not very, if you listen to Tshilidzi Marwala, a South African academic who is now the rector of the United Nations University in Japan and who did his PhD at the University of Cambridge, on precisely this topic.
“AI could negatively impact unemployment in South Africa by causing job displacement, and [a] skills gaps, and exacerbating inequality,” he tells Currency, even if new jobs are being created in other areas, like data science.
The problem, Marwala says, is that AI is likely to take unskilled jobs first – and South Africa’s education system is ill prepared to equip the country with the new digital skills needed to mitigate this impact.
“This, coupled with a largely unskilled workforce, creates a potential perfect storm where AI could exacerbate existing inequalities. AI could take over unskilled jobs, leading to jobs [being lost in] agriculture, manufacturing, and basic services,” he says.
The danger of this scenario, he adds, is that you could have a two-tiered society, with a small, highly skilled workforce benefiting from AI, while many others are condemned to poverty.
So what, then, would be the solution?
According to Marwala, there is no shortcut, other than a “radical education reform”, to provide the country with the right skills in science and technology and digital literacy.
To soften the blow, the country must prioritise reskilling programmes while providing tax breaks for AI-related job creation.
This is because, while AI will automate routine tasks, it will “also create new roles that demand creativity, problem-solving, and emotional intelligence – qualities that are inherently human”.
The immediate implication is that our education system needs to be repurposed to provide not just technical skills useful in careers like engineering, but also critical thinking, creativity and adaptability.
As you might imagine, Marwala is big on education, and in his previous role as vice-chancellor of the University of Johannesburg, he insisted that AI programmes be made compulsory for all students in 2019. He believes other institutions ought to follow suit, pronto.
“The future impact of AI on unemployment in South Africa and globally depends on our choices today. By proactively addressing challenges and seizing opportunities, we can harness it for a more inclusive and prosperous future,” he says.
Birthing a new economic system
But if that is the short-term picture of what is likely to happen to our economy due to AI, what is the longer-term prognosis?
Here, South Africa’s fate will be in lock step with the rest of the world.
In a widely circulated essay this month, Dario Amodei, the CEO of technology company Anthropic, wrote about the dangers of AI under the unlikely title “Machines of Loving Grace”.
“Most people are underestimating just how radical the upside of AI could be, just as I think [they] are underestimating how bad the risks could be,” he said.
Amodei said there are many disturbing elements to this narrative, not least the almost religious fervour of its acolytes who tend to downplay the risks, coming across like blunt propagandists.
AI could be “smarter than a Nobel Prize winner across most relevant fields”, capable of building robots or equipment for its own use.
He rattled off a list of possible advantages from AI over the next 12 years alone, including the elimination of nearly all cancers, extending human life, and eradicating nearly all infectious diseases.
Economically, the benefits are far less clear. As long as AI is only better at 90% of a given job, Amodei argued, it is possible that the other 10% “will cause humans to become highly leveraged, increasing compensation and, in fact, creating a bunch of new human jobs complementing and amplifying what AI is good at”.
But in the long run, Amodei said there will come a point where AI is so cheap and effective that “our current economic setup will no longer make sense”.
Society would then need to reorganise its economy – in the same way that the hunter-gatherer economy gave way to farming – though it’s hard to imagine precisely how right now.
“It could be as simple as a large universal basic income for everyone, though I suspect that will only be a small part of a solution. It could be a capitalist economy of AI systems, which then give out resources to humans based on some secondary economy of what the AI systems think makes sense,” he said.
Or, just perhaps, Amodei argued, humans will continue to find a way to be economically valuable after all. If the robots decide to let us, that is.
Sign up to Currency’s weekly newsletters to receive your own bulletin of weekday news and weekend treats. Register here.
