The adoption of artificial intelligence (AI) in all spheres of life is an unfolding and accelerating megatrend that is greatly changing how we live, work, and socially interact.
A new study done by University of Stellenbosch Business School (USB) graduate of the MPhil in Futures Studies programme, Jan Hofmeyr, found that artificial intelligence can do much more than just improve South African government processes – it can if applied right, improve the reach and quality of services to its citizens.
“In light of rapid advances in the artificial intelligence field, decision-making is gradually being surrendered to machine learning algorithms. The government will in coming years have to concern themselves with the scope and nature of their mandates to regulate and exploit AI for the greater good of societies as they enter an era that is broadly referred to as the Fourth Industrial Revolution,” Hofmeyr says.
Hofmeyr’s research is posed against the backdrop of massive socio-economic challenges with potential political implications and despite just over 27 years of democratic rule, the South African society is stricken by rising levels of poverty and inequality.
“The rapid growth in the range and capabilities of artificial intelligence will in coming years present the government with ever-expanding opportunities to improve the reach and quality of services to South Africans. If they leverage it wisely, this could hold particular promise for South Africa, and other developing countries, to leapfrog stubborn, often structural, developmental challenges,” he says.
He cautions that the adoption will have to be considered alongside ethical governance implications. “Given the wide-ranging concerns around issues of transparency, accountability, and the implications for individual liberty, democracies, but particularly those in the developing world, may be torn between the imperatives for accelerated development and values that are core to liberal-democratic governance.”
Hofmeyr sketches four scenarios depicting where South Africa can be in 2030 when it comes to the adoption of artificial intelligence:
- Everyone’s China: An illiberal democracy with a disjointed approach to its use of an expanding digital ecosystem. It uses AI to expand the services and economic agency of the majority of South Africans, but the ethics that it applies in doing so may fall foul of the constitutional imperatives for transparency and accountability.
- Estonia of the South: A democratic state, with digitally literate citizens, which leverages AI strategically in a transparent, ever-expanding digital ecosystem, by increasing access to services, ensuring resource efficiency within the public service, and enabling individual economic agency for the majority of South Africans.
- Paradise Islands: A democratic state where the benefits of AI only accrue to large corporates and affluent individuals. In the absence of a long-term strategy to expand the reach of digital infrastructure and the government’s technical capacity to make the state more efficient and accessible, the divide between rich and poor is starker than ever before.
- Hola Venezuela: A bankrupt, notionally democratic state, which governs in parallel and sometimes at the order of non-state actors that have filled the governance voids left by its weakness. The internet is censored and AI expenditure prioritises surveillance and predatory technologies, aimed at keeping the desperate masses at bay.
Hofmeyr says the Estonia of the South is the most desirable outcome. “It imagines a situation where values such as individual liberty and transparency and accountability are transferred into the digital sphere. Its focus is to enhance the agency of citizens to reach their potential.”
However, reaching the ideal outcome will depend on various aspects that may influence the use of artificial intelligence as a governance tool in South Africa.
“The resilience of the country’s democracy in the coming decade will depend on the creation of a wide-reaching digital ecosystem, underpinned by a democratic digital ethos. This may somewhat be determined by those that govern, but also by how the government navigates the growing polarisation between the West and China in the digital sphere, which some believe could ultimately result in a divided digital global order,” he says.
Hofmeyr proposes that in response to these challenges, the country may have to prioritise the development of a coherent AI governance strategy that considers questions of infrastructure, but also ethics.
“To reduce the impact of external influence and, hence dependence, it may serve the country well to invest in research and development that focus on the creation of AI technologies that respond to South Africa’s unique developmental needs. It is of critical importance to develop a digitisation strategy that considers the country’s unique needs and interests, or risk being left behind,” he emphasises.
He says liberal democracy in the digital sphere also needs physical access to the infrastructure to enable participation and inclusion of the views of all being governed. “South Africa needs to prepare effectively to compete globally, but also function optimally at a domestic level, amid a growing connection between the physical and digital spheres,” he says. “Underpinning this is imperative for societies to thrive and for individual human beings to live meaningful lives.”
He adds that an artificial intelligence policy that determines the gathering, processing, and optimisation of digitised data, lies at the heart of this endeavour. “Governance by the government will play a central role in shaping the kind of society where the distinction between the physical and the digital will become increasingly blurred.”
He makes the following recommendations on five areas that the South African government can take to ensure a coherent approach to the use of AI that is responsive to its developmental needs:
- Strategy: There is great gains to be made in terms of efficiency in cutting government red tape, improved monitoring and evaluation capacity of policy outcomes, better identification and allocation of resources to priority areas, greater transparency in government procurement, and improved delivery of key services such as health, education, and social development.
- Infrastructure: The Covid-19 pandemic has exposed critical infrastructural vulnerabilities in key areas such as the delivery of health services, education, and the functioning of the country’s social welfare system. A proper audit needs to be conducted of the entire South African digital ecosystem by focussing on its reach, sophistication, and the technologies that are deployed within it. Drawing on the findings, opportunities have to be identified, as well as commercial and regulatory bottlenecks that may detract from their expansion to a broader spectrum of South Africans.
- Research and development: It will be necessary to nurture national expertise in the development of AI applications that can respond to these challenges. The existence of local expertise would be vital to ensure an overreliance on one of the other systems.
- AI Ethics: There is currently no globally enforced standard that regulates innovation within the AI field, which is particularly concerning amid a global race for dominance in this field. In the absence of such normative guidance at the global level, the implication for South African context is t need for vigilance to ensure that new technologies in this sphere complies to the spirit of our democratic constitution.
- Regulation: One of the pivotal points of argument in global debates around the use of AI is the treatment of private data. In South Africa, POPIA (South African Government, 2020), which was promulgated in July 2020, aligns strongly with a citizen-centred approach. Critical, however, will be the extent of enforcement, and the degree of digital literacy required for people to assert their rights.