Amidst daily warnings of an impending artificial intelligence (AI) apocalypse in large parts of the world, South Africans find themselves grappling with more mundane concerns. Geopolitical tensions, economic malaise and inequality dominate our thoughts, leaving little room for contemplating the potential risks and opportunities of AI.

However, dismissing the impact of AI as a whimsical concern would be a grave mistake – a missed opportunity to engage in a vital debate and shape this epoch-defining technology.

While South Africa wrestles with its challenges, the global community has recognised the urgent need to address AI risks. Governments, tech leaders, and experts are engaging in discussions on AI governance, acknowledging its potential maleficent use, biased outcomes and socioeconomic disruption.

The recent calls for AI regulation by the G7, congressional hearings in the US, and the EU’s advanced adoption of AI-focused laws highlight the growing awareness of the need to navigate this evolving landscape. Even developing countries such as Brazil and China are introducing regulatory frameworks.

Optimism overshadows risk

In contrast to the global discourse on AI, South Africa’s response has been relatively subdued. The crucial discussions on AI’s societal role and governance have been largely absent from our national dialogue.

This is not to say that South Africans are not fascinated by AI. On the contrary, the exponential growth in Google searches related to AI over the past year indicates a burgeoning interest. However, the question is whether we are approaching AI with the necessary scepticism, humility and contemplation of its profound potential impact. Recent data suggest we may be falling short in this regard.

A multicountry survey reveals that 57% of South Africans express “trust” in AI systems, surpassing the global average by a significant margin. Similarly, 58% believe benefits of AI outweigh the risks, a figure higher than the median response across surveyed nations.

Furthermore, another survey found an astonishing 78% claim to possess a “good understanding” of AI, ranking South Africa as the most knowledgeable nation among the 28 participating countries, including tech-savvy states like South Korea.

These statistics paint a picture of optimism, relatively speaking, within South Africa regarding AI’s potential, yet also highlight a possible blind spot concerning the risks involved. A striking comparison can be made when contrasting South Africa’s responses to Europe and the US, where 91% of respondents in a survey agreed that AI “requires careful management”.

Locally, much of the anxiety surrounding AI revolves around the spectre of job losses. In a recent survey, a striking 62% of South Africans expressed concern that AI would lead to unemployment, the highest percentage among the 13 countries polled. This focus on jobs is understandable, given the country’s high rate of unemployment, but it also overshadows broader and more far-reaching AI risks.

Nevertheless, there is promise in the fact that South Africans recognise the transformative potential of AI, with 72% believing that it will “profoundly change” their daily lives. This recognition prompts an important question: “How will AI impact South Africa?”  

When reflecting on this, South Africans should be mindful of two factors: firstly, all technology has unforeseen consequences and, secondly, AI is not a public good – it is made by corporations for profit. Examining social media provides a useful point of reflection to explore these factors.

Unpredictability of technology

The path of technological progress is a winding and unpredictable journey. As historian Jared Diamond astutely noted in Guns, Germs and Steel, “technology finds most of its uses after it has been invented rather than being invented to solve a foreseen need”.

This observation highlights the inherent challenge of envisioning the profound and unforeseen uses and impact of emerging technologies from our current vantage point. A prime example lies in the realm of social media.

When Mark Zuckerberg started Facebook, his vision was to “give people the power to share and make the world more open and connected”. Little did he anticipate the intricate web of consequences that would ensue from his invention, including political manipulation, mental health problems, the rampant spread of misinformation, and even real-world violence.

Social media also illustrates how quickly software-based technology can spread and entrench itself in our daily lives. Merely two decades ago, the vast majority of social media platforms we now take for granted did not exist. Yet, today, these platforms boast an astonishing 4.8 billion users, exerting influence over every facet of our lives, both online and offline.

However, the speed at which AI is expanding its reach eclipses even the remarkable growth of social media. A prime example is ChatGPT, which garnered a staggering 100 million active users within only two months. In contrast, it took popular platforms like TikTok and Instagram nine and 30 months, respectively, to achieve the same result.

What does a world look like where nearly 5 billion people, both benign and malicious, have easy access to a range of AI applications? We simply have no idea, although some of the forecasts are alarming.

For profit, not prosperity

Social media platforms have emerged under the control of profit-driven companies that compete against each other for our attention. These applications were not conceived to operate with societal interests at the forefront, but rather with the ultimate aim of maximising revenue.

The profit motive behind social media, coupled with a competitive system and lack of regulation, has yielded a range of destabilising effects on individuals, institutions, systems, and societies. Furthermore, these companies are accountable to shareholders rather than the public, raising concerns about their alignment with the greater societal good.

It is important to clarify that this does not imply moral repugnance on the part of social media firms or denounce profit generation as inherently wrong. However, in the absence of robust regulation, an inherent incentive problem arises, potentially leading to harmful behaviour.

Similarly, the field of AI faces a regulatory void and is primarily driven by corporate entities. Until 2014, academia predominantly released machine-learning systems, but industry has since taken the lead. In 2022 alone, 32 “significant” machine-learning systems – the backbone of most AI models – were produced by industry, in contrast to a mere three by academia.

This trend is poised to continue, as the development of cutting-edge AI systems requires substantial quantities of data, computing power and financial resources – areas where industry actors possess distinct advantages over non-profits and academia.

Right here, right now

While we cannot foresee AI’s exact impact, we do know that fast-moving technology comes with widespread consequences, which, overlaid with unregulated profit-seeking by competing firms, can result in suboptimal societal outcomes.

However, we must resist the notion that this result is predetermined – technology does not dictate destiny. South Africans are right to be optimistic about AI and fully embrace its positive potential.

Yet, in doing so, we must also assume responsibility for shaping our own future and critically assess the country-specific risks and consider mitigation measures.

AI may feel distant from our immediate problems, but the technology is here, whether we are ready or not. South Africa can either join others in debating and shaping the trajectory of AI – or be shaped by it. DM