Why do we need an AI policy in South Africa?

Research ICT Africa’s researcher, Alex Comninos asks, “Why do we need an AI policy in South Africa?”

Comninos delivered this input as part of a panel discussion at the Bo-Kaap Cultural Hub in Cape Town on 10 March 2020 at the launch of the Human Sciences Research Council’s Topical Guides Series: AI and Data.

The context that Comninos used to frame his input is, firstly, that South Africa (SA) is still struggling with significant aspects of the third industrial revolution – our electricity crisis being a case in point, let alone the challenges of the Fourth Industrial Revolution (4IR).

Secondly, while artificial intelligence (AI) does highlight new challenges for us, SA is still struggling with issues such as access to ICT. For example, RIA’s After Access Surveys find that only 53% of South Africans are online.

He questions the need for a specific AI policy framework and ethics guidelines – arguing that there are currently over a hundred documents on ethics, norms and principles on AI. He argues that it would be better for us to take a step back in light of the fact that we already have a human rights framework that gives us ethical direction, norms as well as principles that we could build upon to apply to AI.

One of the dangers of an AI specific set of policies is that it is led by the private sector and is often voluntary. In this regard, we can lose focus on areas of human rights regulation where we can build our existing capacities.

Also on the panel were Varsha Sewlal (Information Regulator) and Fadlah Adams (South African Human Rights Commission). The panel was facilitated by Dr. Rachel Adams of the Human Sciences Research Council (HSRC). Dr. Michael Gastrow (HSRC and Presidential Commission on the Fourth Industrial Revolution) was the discussant.

Related