Inside Africa: The AI Bandwagon’s Here: Wait, Nevermind: South Africa’s Draft National AI Policy Withdrawn
Matheus Bertelli
On April 10, 2026, South Africa’s Department of Communications and Digital Technologies (DCDT) officially published its draft National AI Policy, one that could potentially address the unequal development throughout the country if executed accordingly. However, it was withdrawn following increasing allegations that the draft itself was written using AI. But before examining the controversy, it is worth considering what the policy stood to offer before its withdrawal.
What the Policy Stood to Offer
In the draft AI National Policy, the South African government laid out its main aims clearly, which encapsulate the national goal of establishing South Africa as the continental leader in pioneering AI adoption. Domestically, the goal ranges across two pillars of ethicality and development. On the ethics pillar, the government set out to ensure the ethical governance of AI systems, the development of oversight infrastructures for the ethical use of AI, sustainability and inclusivity, and intergenerational equity in AI innovation. Regarding national development, the draft listed goals to build AI digital infrastructure and supercomputing facilities, as well as to embed AI use in education.
The policy targeting AI integration in education, in terms of both classroom content and the AI’s technical integration, is a significant policy step for South Africa; if implemented, the policy has the potential to mitigate the lack of quality in education resources across different regions and the lack of AI talent pool in South Africa.
One of the key struggles in South Africa’s strive for greater educational quality is that often teachers are not equipped with proper teaching qualifications, and that students lack access to quality teaching materials. The integration of AI into education would enhance the learning process through personalized tutoring, which has been proven to be more effective than overcrowded classroom learning, thereby filling in the gap left by poor-quality educational resources.
Beyond encouraging AI use, the South African government aimed to bridge the AI skills gap, equipping the young generation with the highly demanded proficiency in the use of AI in their future careers. In fact, several banks across Africa are already distributing co-pilot licenses to their employees for productivity gains. Therefore, ensuring students are well-equipped to utilize these LLM tools will help in improving their job prospects,in this case within the banking industry, which would help in socioeconomic advancement.
Where the Ground Reality Bites
In section 4.2 of the draft policy, the South African government stated the use of AI for rural development as one of its objectives. However, the blueprint for AI usage across different sectors remains unclear through the draft itself. What is clear is that South Africa remains deprived of the infrastructure needed to implement the use of AI across rural regions. As acknowledged in the draft itself, the implementation of these policies rely on “robust digital infrastructure” and “widespread connectivity”, which remain obstacles for South Africa at this point in time. The fragmented access to digital infrastructure across South Africa is compounded by financial, literacy and geographical barriers, causing unequal access to digital infrastructures especially in rural regions like Mkatazo village.
According to research, Mkatazo village is one of the rural areas in South Africa whose population is digitally excluded due to the lack of proper ICT infrastructures such as fiber optic internet services; their digital connectivity coverage is more than 10% lower than urban areas. Digital literacy remains poor, with students excluded from digitized educational resources and business owners prevented from expanding their businesses due to the lack of digital integration in their business models. Mkatazo village is not alone; it is a microcosm of the digital deprivation of most, if not all rural regions in South Africa. In this context, there is a risk of the National AI policy becoming a tool that exacerbates inequalities within those structurally and economically positioned to benefit from it and those who aren’t.
A Reason for Optimism
A point of optimism worth noting is that Eskom – the primary provider of electricity in South Africa, has celebrated its 300th day without an unplanned blackout on the midnight of March 12, 2026, a concrete sign of improvement on the recurring load shedding phenomenon in South Africa. It indicates a possibility that one of the most critical chokepoints – load shedding – in the nation could be permanently removed if this progress is sustained, which would ease the implementation of nationwide AI usage, especially in areas like healthcare or public services, which the government has mentioned in the draft.
Withdrawal of the Draft AI National Policy and What This Means
Despite extensive discussion on the positives and negatives of the policies and the government’s call for the nation’s input on the draft, it has since been withdrawn following its release on April 10, 2026, ironically due to the suspected use of AI. The withdrawal of the AI policy itself serves as an artifact highlighting the wider challenge of the ethical use of AI and the potential harm of overreliance on AI in the professional sphere.
Nonetheless, the government's ambition in hopping on the bandwagon of AI is real; the challenge of building equitable digital infrastructure is a steep one, but one that is progressively improving, as seen in Eskom's 300-day streak without load-shedding. Whether the government would revise a new AI national policy and whether the new draft would address elements of inclusivity remain to be seen, but one thing is clear: South Africa has missed the bandwagon to implement nationwide AI regulation, and if South Africa wants to be on it, the governmental departments need to substantiate each claim before the next release – until then, we're patiently waiting.