Trust does not sustain itself; this was the opening premise of a panel I participated in at the inaugural LexisNexis Future of Law Conference in Sandton last week, under the theme “The Architecture of Trust: Accountability in the Intelligent Age.” Organisers noted that trust depends on structures—codes of conduct, governance frameworks, disclosure obligations, professional standards, and institutional culture—that define accountability. The key question is whether this architecture remains fit for purpose now that artificial intelligence is embedded within the governance structures of the institutions it was designed to oversee.
My view, which I will elaborate below and suspect may not be universally shared, is that the current architecture is fundamentally sound. While it is fragmented and, in some areas, lagging behind, it is not, in principle, deficient. The principal shortfall in many South African boardrooms is not the absence of legislation; it is a lack of visibility.
Directors cannot delegate their responsibilities to machines. Our current framework is fragmented but not lawless, and while it is outpaced in practice, it remains sound in principle. South Africa has experienced rapid regulatory cycles, first with COVID‑19 in 2020 and then in late 2022 to avert placement on the FATF grey list. Both sets of regulations were issued with urgency, and at least one clearly achieved its goal: South Africa is no longer on the grey list. Although the merits of the COVID‑19 rules are debatable, both pandemic response and AI share a common challenge: rapid spread and swift evolution. Early pandemic rules struggled to keep pace with the virus’s mutations, and similarly, AI regulations must avoid being locked into frameworks designed for earlier “versions” of the technology.