Abstract
The twenty-first century is experiencing the exponential growth and integration of artificial intelligence (AI) across the global business landscape. AI offers companies unprecedented benefits but also poses significant risks. While AI adoption in South African companies is accelerating, it is occurring amid dire socio-economic conditions and in an unregulated environment. This study examines whether South Africa’s corporate governance and regulatory frameworks are adequately equipped to manage these risks.
These risks posed by AI translate into risks for companies, making decisions to deploy AI systems a board responsibility. Boards are mandated by section 76(3) of the Companies Act to exercise their fiduciary duties in the “best interests of the company”. They are responsible and accountable for proactively overseeing AI systems and ensuring that the appropriate risk mitigation strategies are in place. The irresponsible deployment of AI would have devastating consequences for South Africa’s fragile labour market, potentially harming companies’ long-term financial stability.
This study identifies a significant governance gap. Current legislation, such as the POPIA, ECTA and Cybercrimes Act, is reactive and inadequate to address the challenges posed by modern AI. Corporate mechanisms such as the Social and Ethics Committee (SEC) lack the statutory authority to monitor technology and artificial intelligence. In its pursuit of AI-dedicated legislation, the South African government can learn from the European Union’s Artificial Intelligence Act (AIA), but this study advises against indiscriminately copying from it.
The study culminates in a proposal that, in the absence of AI-specific legislation, companies should self-regulate via an Internal Regulatory Framework (IRF). And it recommends amendments to section 72(4) of the Companies Act, designating the SEC as a board committee and to regulation 43(5), incorporating the King Code and empowering the SEC to monitor technology and artificial intelligence. These proposed changes would provide the SEC with the legislative authority to fulfil the function of an internal AI regulator, thereby ensuring that AI deployments by companies are subjected to proper oversight and management.