South Africa’s proposed AI plan needs a rework: experts
South Africa has taken its first steps in establishing a national AI policy and subsequent regulatory framework by recently publishing its Draft National Artificial Intelligence (AI) Plan discussion document.
However, various technology and legal experts have critiqued its contents, saying that the document does not adequately prepare the country for the impact of AI.
The document “appears to have been published as a rough draft – it is repetitive, has conflicting provisions and is not sufficiently advanced, specific or practical in clarifying and setting a clear policy approach and informed plan to offer meaningful guidance on the way forward,” said Werksmans’ head of technology and innovation Preeta Bhagattjee and senior associate Savanna Stephens.
The Department of Communications and Digital Technologies published the Artificial Intelligence (AI) Planning Discussion Document on 5 April 2024.
Bhagattjee and Stephens said that the aim of this is to encourage a dialogue between the public and private sectors in order to promote innovation in AI and to develop a national AI policy.
This policy would outline the government’s stance on the integration of AI in South Africa, government-led AI initiatives, and proposed principles to guide the use and development of AI in the country.
The Discussion Document provides early insights into the government’s strategy for AI adoption, with key themes including:
- A national strategy for AI is needed to advance economic growth;
- Develop and implement regulatory approach to AI in South Africa over the period of 2025 to 2027;
- To address ethical issues, it’s crucial that the adopted legal framework or national AI policy includes guiding principles;
- To address future AI skill needs, the government should invest in initiatives that enhance AI literacy through education and training, support local AI solution development and improve public data systems in key sectors for better e-government services using AI.
Director at PPM Attorneys, Lucien Pierce, wrote that the draft is “not good enough” as it is “convoluted, complicated, lacks clear deliverables and fails to allocate responsibility for their delivery.”
Pierce said that the draft is overly lengthy, jargon-heavy, and poorly structured, with numerous errors and “seems to have unfinished thoughts,” and outlines a vague and high-level AI adoption plan, without assigning specific responsibilities or detailing how to achieve its ambitious goals.
Bhagattjee and Stephens echoed this, saying that “it is largely styled as a list of options, initial thoughts and information from other sources and records of developments in AI use and governance and regulation in other countries.”
Additionally, Pierce said that it introduces unclear concepts like “Centralised Computing Power” and proposes “unrealistic” timelines for implementing AI regulation and infrastructure projects.
While the draft draft suggests a degree of data sovereignty to manage the extensive data needed for AI training in South Africa but fails to mention the essential Draft National Data and Cloud Policy or address the challenges in securing the computing power needed for its objectives, according to Pierce.
“My view is that, rather than stakeholders commenting on this highly flawed Draft AI Plan, it should be completely reworked and released when it is in a more practical and improved form,” Pierce said.
Positively, Bhagattjee and Stephens said that the Discussion Document indicates progress towards establishing AI policies and legislation in South Africa.
They anticipate that an updated version will be released, incorporating feedback from essential stakeholders across various sectors and AI organisations.
The experts say that the reworked version should outline clear strategies for South Africa, drawing lessons from international practices to both encourage AI development while addressing associated ethical and harm risks effectively.
Read: South Africa requires comprehensive AI legislation – experts