SIYABULELA MAKUNGA | SA capable of build Al sovereignty

As digital markets continue to expand and artificial intelligence (AI) evolves, governments and regulators alike need to remain agile when it comes to governance and public strategies in AI.

Just last week, the deputy minister of communications and digital technologies Mondli Gungubele led the official launch of the National Artificial Intelligence Stakeholder Forum in Pretoria, to mark “a decisive step in SA's mission to build an inclusive, ethical, and people-centred digital ecosystem.”

The forum brought together key role players from across the AI value chain including researchers, engineers, innovators, ethicists, public servants, academics, and private sector leaders in order to co-create a shared vision for SA's AI future.

“Every subcontractor from the rural child to the urban entrepreneur must be part of this technological revolution,” said Gungubele.

These prevailing AI regime places a greater value to the competition policy as an important tool that when effectively applied can regulate adverse effects from the operation of AI and Big Tech.

One of the commission’s senior analysts, Donnavan Linley, recently participated in a panel discussion on AI governance and economic inclusion at a pre-Brics summit in Rio de Janeiro, Brazil, where he highlighted three major points central to effective regulation and opportunity creation within the evolving landscape of AI.

Every subcontractor from the rural child to the urban entrepreneur must be part of this technological revolution
— Deputy minister of communications and digital technologies Mondli Gungubele

First, market inquiries are a regulatory or competition policy tool that can address AI’s impact on markets and consumers. With AI and AI-embedded digital platforms in constant evolution, market inquiries provide a flexible, evidence-led mechanism that involves stakeholders in seeking practical solutions.

This approach supports the design of remedies that mitigate the adverse effects AI might introduce, without constraining the innovation and benefits that AI products and services may offer.

Securing data sovereignty was one of the summit’s key discussion points as delegates also unpacked issues around access to data. Access to data especially in training AI models such as large language models (LLMs) are necessary as LLMs require a large scale of content and data to develop.

Currently Big Tech has the infrastructure required to access content and data. However, this creates high barriers to expansion for small and medium enterprises (SMEs) and start-ups facing significant challenges accessing essential data that will enable them to further develop their AI models.

There is now a critical need for the inclusion of South African data and data sharing in the development of LLMs as it is vital to add that local context and responsiveness to AI chatbots. If we act now as regulators, we can position competition policy to address the emerging data asymmetry between Big Tech platforms and local South African companies.

The third point our analyst raised relates to the access to AI infrastructure. We cannot deny that building digital infrastructure for AI (i.e. data centres, supercomputing, high-speed networks, etc.) is capital-intensive. Competition policy can play a role where access to AI infrastructure (e.g. cloud services) must be fair, non-discriminatory and non-exploitive.

In the past, regulators might not have been synonymous with agility and collaboration. Through market inquiries, in particular, the Competition Commission has proven that it is agile and open to working with firms and industry stakeholders to find solutions that address competition concerns, create opportunities for SMEs and start-ups, and balance the rights of consumers.

Not only is there scope for SA to develop its own AI sovereignty, but opportunities also exist for the development of niche and industry-specific AI capabilities such as AI applications and LLMs.