Presently, the South African legislative framework makes provision for protection of personal data – but this framework seeks to protect and entrench data only as far as the use and processing of sensitive or personal information is concerned.
The draft Open Data and Cloud Policy introduced by the Department of Communications and Digital Technologies is a step in the right direction; it seeks to strengthen the capacity of the state to deliver services.
An AI maturity assessment goes wider; it seeks to understand existing systems, determines the areas of weakness in the systems and provides relevant evidence to support efforts to promote the use, development, and deployment of AI systems holistically.
Evidence-based policy provides an opportunity to create policy that benefit and not hinder technology development and innovation. Policies that directly govern AI-based technologies as well as those that indirectly apply to broader AI fields, require in-depth data analysis and use cases to fully understand the risk, benefits, and opportunities, before being legislated.
Areas immediately affected by AI include health, education, automotive manufacturing, mining, and financial services – all key sectors in South Africa. However, governance of AI in each individual sector is complicated by several factors and it would be unfeasible to have policies for individual sectors. However, cooperation with and key data from these sectors can be used to create a central framework that establishes categories of risks and how use cases in each category could be regulated. This standard of assessment can then be measured against other countries or regionally, to benchmark the assessment, with a view to adopting or creating regional AI ecosystems.
The AI Maturity Assessment Framework (MAF) will help decision-making stakeholders in South Africa to determine:
The framework will consider best practices globally and locally, including existing frameworks and indices, but be adapted specifically for the South African context – it would illustrate an AI benchmark to measure use of data and AI resources in South Africa in comparison to the other countries in the region.
The framework will also be supported by a dynamic dashboard tailored for this initiative, that would visualise the progress being made over time, through monitoring of the indicators.
The expected outcomes of the above outputs are as follows:
March 2023 – annual AI MAF reports