Graphic: John McCann
The governance of artificial intelligence (AI) involves a complex balancing of dilemmas — the trade-offs between opportunity-seeking and risk aversion, the interplay between security and transparency, tension between globalisation and localisation and the interplay between self-regulation and government control.
Indeed, AI governance frameworks must balance innovation, ethics, public safety and fairness as technologies evolve. Flexible, inclusive, context-aware and predictive AI governance policies are needed. Multi-stakeholder cooperation between governments, commerce, academia and civil society is essential to guarantee robust and adaptable regulatory frameworks.
What is critical is a dynamic and balanced governance structure that can adapt to AI advances to ensure the technology delivers maximum benefits with minimum risks. South Africa is no exception, but we have fallen woefully behind in our response. We thus make the argument for a new model that considers this rapidly shifting landscape alongside the need for legislation.
A dynamic governance framework is essential
It is acknowledged that AI is a fast-changing technology, with innovations sprouting all over the world. If we consider the global context, a dynamic governance framework becomes crucial for AI adoption because it allows regulatory structures to adapt to rapid technological changes, diverse applications and the evolving societal impacts of AI.
There is a fundamental need for adaptability to rapid change. A dynamic governance framework can evolve alongside these advancements, ensuring regulations remain relevant without stifling innovation. There is a need to balance innovation and regulation. Dynamic governance enables parity between encouraging AI development and ensuring responsible use. Instead of rigid one-size-fits-all rules, adaptable governance frameworks can implement risk-based regulations, differentiating between low-risk and high-risk AI applications and adjusting requirements as the technology matures.
AI raises various ethical and social concerns around bias, privacy and accountability. A flexible governance model can adapt to these issues as they arise, allowing for ethical considerations to be incorporated into policy over time rather than being locked into initial, potentially outdated standards.
AI development and deployment are global, so governance frameworks must accommodate different legal, social and cultural contexts. A dynamic approach allows for international collaboration, helping to harmonise standards and reduce regulatory fragmentation across borders, making it easier for AI systems to be adopted globally. As systems increasingly impact society, engaging a wide range of stakeholders — developers, users, governments and civil society — is essential. This structure enables ongoing consultation and input from stakeholders, making policies more inclusive and grounded in real-world applications and concerns.
AI carries specific potential risks, including cybersecurity vulnerabilities, job displacement and unintended consequences from autonomous systems. A dynamic governance model can respond to these risks with targeted measures, such as evolving security standards and adaptive workforce policies, to address risks as they become better understood.
Public trust in AI technologies is essential for their widespread adoption. A transparent governance framework that adapts to address new risks and concerns can help build confidence, ensuring that AI systems are seen as safe, fair and beneficial.
It must also be noted that inaction poses its own kind of risk.
Towards a dynamic technology law commission
We argue that a dynamic governance framework for AI allows policymakers and stakeholders to respond flexibly to new insights, ensuring that regulations evolve in tandem with technology to support safe, effective and ethical AI adoption. In particular, a dynamic technology law commission, consisting of tech-savvy lawyers and AI experts, can help achieve dynamic governance. Establishment of such a structure must be pursued in all jurisdictions, worldwide.
To home in on the South African context, in 2020, following an extensive process, the Presidential Commission on the Fourth Industrial Revolution (PC4IR) report was released. It called for: “Radical innovation to deliver the required speed in enacting legislation which enables extraction of maximum economic and societal value.”
Five years on, the “required speed” has actually been rather static. Historically, the process of crafting and adapting legislation has been a slow, and often painful, process. But as the world quickly changes in the 4IR, so too must our approaches.
Prior structures simply did not anticipate that machines would learn how to mimic our intelligence or that decisions about life and death could be delegated to algorithms. With these considerations, Letlhokwa George Mpedi, vice-chancellor and principal of the University of Johannesburg, has long argued for a shift from reactive legislation to anticipatory legislation. As the American legal scholar Lawrence Lessig argues: “Law and technology produce, together, a kind of regulation of creativity we’ve not seen before.”
If we are to heed the advice of the PC4IR report then the seventh recommendation calls for us to review, amend or create policy and legislation. It makes a compelling argument for parliament to look at South African legislation and update it in line with the 4IR. Accordingly, the recommendation also calls for legislature and state executives to be 4IR trained to implement the necessary changes. Perhaps the glaring question then is, “How do we now proceed?”
There is an argument to be made that draws from existing legal structures. There is already a South African Law Reform Commission (SALRC) tasked with making well-researched recommendations for the development, improvement, modernisation or reform of the law. The SALRC was established by the South African Law Reform Commission Act 19 of 1973. The aim is to ensure accessible and just legal systems. The commission places strong emphasis on public consultation to ensure that the law reform process is both participatory and transparent. This commission has largely been considered successful.
The case for a special commission on the 4IR and law
Many of the SALRC’s recommendations have led to important legislative reforms and its research outputs are widely used by legal scholars, courts and policymakers alike. Yet, in our current context, it could be argued that the 4IR is simply too broad to fall under it. In fact, it could be argued that a special commission dedicated to the 4IR and law reform is wholly necessary.
The SALRC, while competent and historically impactful, is structured to address discrete legal issues through a traditional programme-based model that might not necessarily be agile or interdisciplinary enough to keep pace with the rapid global shifts brought about by the emerging technologies that characterise the 4IR.
That is to say that, when the SALRC was formed, we could not have anticipated that this would be our world. The pace and nature of these changes call for a fundamentally different model of legal foresight and responsiveness. This model must consider continuous technological monitoring and multi-sectoral collaboration. A special commission dedicated exclusively to 4IR and law reform is thus not only justified but urgently necessary.
This commission would bring together legal scholars, technologists, ethicists, economists and policymakers to anticipate and respond to the complex legal and regulatory challenges posed by technological disruption. Such a body would be better equipped to craft adaptive and forward-looking legal frameworks that promote innovation while also safeguarding constitutional rights and social equity.
While this body would exist in tandem with the SALRC, it could operate with the speed, adaptability and breadth of expertise required to ensure that South Africa’s legal system considers accelerating technological change and its implications.
The Greek philosopher Heraclitus once said: “There is nothing permanent except change.” Though the ancient Greeks did not anticipate the world we would live in centuries later, this thought certainly endures. If we consider Heraclitus’ words through the lens of the legislation, then it stands to reason that societies and the law that govern them must evolve with the times. Now, that includes a consideration of how laws are formed. A special commission on 4IR and law reform could thus be considered a necessary institutional evolution — and one we dare not miss.
Letlhokwa George Mpedi is the vice-chancellor and principal of the University of Johannesburg. Arthur Mutambara is a full professor and director of the Institute for the Future of Knowledge at the university.