Al plan could protect SA’S economy – and people

The technology has huge implications for the country and the world. Experts say if South Africa fails to prepare, it will be preparing to fail

The South African government has started drafting a plan in preparation for what looks to be an economy-shaping artificial intelligence (AI) boom. And, although some have already criticised this first step, others are relieved that the government is making the effort. Because what is clear is that, without planning, the country’s economy — and people — will be left vulnerable, while also missing out on a game-changing opportunity.

In late March, the department of communications and digital technologies published a 53-page discussion document outlining South Africa’s current AI landscape, the pitfalls and opportunities presented by this technology, as well as timelines for adopting a plan.

The document — which was slammed by some commentators for failing to present a coherent approach to AI policy — envisions that South Africa will develop a regulatory framework between now and 2027. Among other things, the document proposes that the government invest in locally developed AI solutions.

In drafting the plan, South Africa is following the lead of a number of other countries, including members of the European Union, which passed the world’s first comprehensive AI law in March this year.

In April, Nigeria’s government drew on the expertise of 120 researchers and practitioners to develop a national AI strategy, which has reportedly already attracted $3.5 million in seed funding. The development gives Africa’s biggest economy a head start in the race for AI innovation and regulation on the continent.

Meanwhile, AI is at the centre of the unfolding Cold War between China and the US. Just this week the latter country’s authorities announced new steps towards managing Ai-related risks. And through its strategy, initiated back in 2017, China has ambitions of becoming the centre of AI innovation by 2030.

Remarking on the need for South Africa to implement its own plan, Leanne Mostert, a partner at Webber Wentzel and a commercial intellectual property law expert, said: “AI can be used in ways that we have never anticipated. It can be used on a wider scale than has ever been anticipated.”

In creating a plan, countries such as South Africa will have to balance the risks associated with AI — including the spread of disinformation through deepfakes and the like, as well as data breaches — against what Mostert describes as “the enormous positives” of the technology.

Last April, Goldman Sachs made a bullish proclamation on the positive economic effects of breakthroughs in generative AI, such as CHATGPT.

According to the investment banking giant, as tools like these work their way into businesses and society, they could drive a 7% increase in global GDP and lift productivity growth by 1.5 percentage points over a 10-year period.

But the positives aren’t only economic.

Also last year, an article published in peerreviewed science journal Heliyon noted that AI may play a key role in accelerating drug design by avoiding costly timeconsuming laboratory work.

But, if patent legislation does not adapt to AI’S emergence, these discoveries won’t be protected. Current legislation protects human-created innovations.

“It is so important that we come to some sort of landing on how we’re going to approach these things, because there are going to be huge advancements in drug discovery and things like that,” Mostert said.

“And we can’t protect these advancements, who is going to put the money in to take those drugs through trials when it’s an enormously expensive process to take them to market? … So it is going to be very important to evolve our legislation or create new legislation to deal with these questions, which have never been raised before.”

The government’s discussion document underlines the importance of attracting investment — something the country has struggled to do in the midst of regulatory uncertainty.

This investment strike could extend to AI advancements, given the hurdles that exist in commercialising South African-owned intellectual property.

“Until we can improve that whole framework … then we have not created an environment that fosters this development,” Mostert said.

“The draft strategy says we must immediately support AI start-ups and create an innovation pipeline. And I agree. But I’m not sure that we don’t need to urgently address those other things if we are ever to be a place that attracts these types of skills and businesses.”

AI ethics policy researcher Emma Ruttkamp-bloem agreed that South Africa needs an AI strategy, saying: “It is absolutely imperative to have a plan … This is a technology that affects every sector from agriculture to fintech. It has immense potential for good. It can help developing countries to progress much faster. But it comes at a price if one isn’t prepared for it.”

Over and above some of the more commonly understood risks, AI has the potential to exacerbate existing inequalities given its tendency to repeat existing data patterns. “It doesn’t have the ability to weigh up consequences. It will simply latch onto a pattern and generate outcomes based on that,” Ruttkampbloem said.

“So in South Africa, which has huge social inequalities, it will be disastrous if certain social groups are excluded from access to education or from receiving mortgages. For us, that is the biggest negative consequence.”

In addition to addressing potential holes in South Africa’s regulatory framework, the country will also need to improve its infrastructure, Ruttkamp-bloem added. The country’s failing energy infrastructure is an obvious hurdle to AI innovations.

“There is a lot of potential in South Africa, but none of that can be fully realised if there isn’t a plan of some kind,” she said.

Remarking on the type of effort that will have to go into developing and implementing such a plan, Ruttkamp-bloem said: “It’s not something that can be done overnight.”

“To just get legislation set up can take a very long time. But you also have to take into account that the different social, political and ethical concerns around the deployment of AI technology arise in different sectors,” she added, suggesting that South Africa would need a dedicated department that focuses on AI.

In the face of the country’s many pressing problems — such as energy and water insecurity — the state doesn’t have the luxury of neglecting to address Ai-related issues, according to Ruttkamp-bloem.

“This is not a technology that we can stop. It is not as if one can choose not to engage with this technology. Because this is how the world works now,” she said.

“And also, it has the potential to solve basic issues … There are a lot of really pressing aspects of ordinary life in South Africa that can in fact be solved faster if we have the infrastructure and engage responsibly with AI.”

‘It doesn’t have the ability to weigh up consequences. It will simply latch onto a pattern and generate outcomes based on that’