National AI Government Summit Discussion Document

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING
Disclaimer – Content, Usage of the Draft Document:

This draft document is for the general discussion purpose and not for publication or academic use. The legal
conditions on the use of some quotes borrowed from other sources related to Artificial Intelligence (AI) is
placed currently as to solicit opinion of participants on their relevance. Some these quotations make it into
the final draft measures would be taken to demonstrate or indicate correct sourcing and acknowledgement
once is about to be made an official document. Therefore, seeking legal counsel before acting based on the
contents of this draft is highly recommended. The omission of trademark or service mark does not imply a
lack of intellectual property rights associated with the mark or logo. All rights reserved as this draft is for
policy formulation discussion purposes.
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT

Prepared by
Alfred “Ali” Mashishi: Chief Director – ICT Forecasting and Modeling

On behalf of the
Department of Communications and Digital Technologies (DCDT)
October 2023
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT i

Foreword by Minister of Communications and Digital Technologies
There are strong indications that AI has the potential to deliver more benefits to humanity as compared to
any technology introduced in the last century. As with any technology, these benefits are always matched
by the ills or harms they carry. The harms requires that they be mitigated and managed through a
sustainable planning process at all levels of society.
Since the invention of computers or computation machines, their growth has been exponential both in
terms of capability and task performance. Our understanding is that this was because of humans’ ability
to have developed the power of computer systems with diverse working domains, speed increase, and
size reduction and time adjustment.
The global assessment on how countries are addressing AI as a general-purpose technology differs in
terms of continental, regional and national circumstances. Various permutations range from seeing AI as
a tool that will bring goodness to humans as outlined by the International Telecommunications Union (ITU),
or AI being able to bring prosperity for humans’ rights according to the United Nations AI Task Force, or
AI challenging human ethical foundations. Ours as South Africa is guided by the Presidential Commission
on the 4th Industrial Revolution (PC4IR) report which looks at AI as a technology tool that will advance our
social and economic prosperity.
We have witnessed several AI I initiatives around the country emanating from the government, private
sector, academia, and civil society and decided that the best approach is to devise a plan that would
encompass all. The plan purpose is to ensure that all the AI key role players can agree on the national
priorities and objectives that will ensure our prosperity and growth through the adoption of AI.
From the government perspective, two (2) major initiatives are already taken and have been widely
accepted. The Department of Communications and Digital Technologies (DCDT) has established an AI
Institute of South Africa (AIISA) in 2022 and the Department of Science and Innovation (DSI) has established
the Centre for Artificial Intelligence Research (CAIR) in past years.
South African government is now outlining key government AI priorities and putting up some deadlines
that would ensure that we are not left behind on the AI global race. We understand that this phase of AI
will grow in leaps and bounds and we will always be guarding against the negative aspects and harness
the positive one and make the African continent a success story.
Mr Mondli Gungubele, MP
Minister of Communications and Digital Technologies
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT ii

Executive Summary
The advancement in Artificial Intelligence (AI) as a technology has created a breeding storm and
necessitates that each sovereign country to have a deep look at its propensity, impact and pitfalls towards
its citizens and needed governance systems. The AI National Plan approaches the computing system in a
logical flow that would provide impetus for adoption and provision of direction to all stakeholders.
The plan begins with the provision of an overview of AI landscape and how is viewed from the perspective
of South Africa’s landscape. Section A outlines the global AI discussion as influenced by trends and drivers.
Global discourse is influenced by political, economic and militarization of AI. It further points out that this
discourse brings out global concerns and requires tools and applications. The regional and national
approaches are assessed and aligned with the South African conceptual framework.
Section B looked at how the South Africa’s landscape has been impacted by the pitfalls and opportunities
brought by both versions of Applied and Generative AI. It unpacks the PC4IR approach and indicate how
AI is impacting South Africa through law, governance, and regulatory approaches. It further elaborates by
sharing localization of the AI approach that will drive South Africa’s aspirations.
Section C outlines South Africa’s AI plan priorities and key enablers. It further provides how the AI
governance and institutional mechanisms could be guided by the ethical and regulatory approaches. The
governance structures to manage AI in South Africa is proposed with the incorporation of AI, Big Data and
IOT (automated services). This is directed by the AI framework which was developed in the aftermath of
PC4IR report approval. It then points out the key priorities to be adopted and the timelines that are within
the outlined AI opportunities.
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT iii

CONTENTS
Foreword by Minister of Communications and Digital Technologies ....................................................................... ii

Executive Summary................................................................................................................................................................ iii
Section A: Modelling Artificial Intelligence (AI) within the South African Landscape ......................................... 1
A1.1 Introduction and Purpose................................................................................................................................. 1
A1: 2 Global Discourse: Politics of AI; Economics of AI; and Militarization of AI......................................... 2
A1: 3 Investment in AI Technologies and Applications....................................................................................... 7
A1: 4 Global Concerns on AI ..................................................................................................................................... 9
A1: 5 Main Types of AI Discourse ............................................................................................................................ 10
A1: 6 Components of AI- AI System; Machine Learning; Neural Networks; Deep Learning.................. 11
A1. 7 Regional vs National AI Approaches and Frameworks .......................................................................... 14
A1: 8 South Africa’s AI Conceptual Framework and Approach ...................................................................... 15
Section B: AI – South African Landscape, Opportunities and Pitfalls ..................................................................... 16
B1. AI Understanding within South Africa’s Context ...................................................................................... 16
B1. 1 Adoption of AI: SA AI Approach PC4IR ...................................................................................................... 17
B1. 2 Applied and Generative AI: Impact in South Africa ................................................................................ 18
B1: 3 State Considerations in terms of Law, Governance and Regulation................................................... 19
B1: 4 South Africa’s Aspirations: Global and Continental ................................................................................ 22
Section C: South African AI Plan ...................................................................................................................................... 23
C1 Priority Focus Areas and Enablers ............................................................................................................... 23
C1. 1 Discussion on Enabling Environment ......................................................................................................... 26
C1. 2 Focus Areas: What must be in place for AIISA to be Effective and Realistic ................................... 32
C1. 3 Data, AI and IOT ............................................................................................................................................... 37
C2 AI Governance and Institutional Mechanisms ......................................................................................... 38
C2: 1 South Africa’s AI Ethical Considerations .................................................................................................... 38
C2: 2 SOUTH AFRICA AI REGULATION APPROACH......................................................................................... 40
C2.3 SA AI and DATA Governance Structure ...................................................................................................... 41
C3 AI Plan Adoption and Timelines Processing............................................................................................. 43
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT iv

Section A:
Modelling Artificial Intelligence (AI) within the South African Landscape
AI The International Landscape: AI Global Discourse, Trends and Drivers
AI “is a software that is developed with one or more of the techniques and approaches (…) and can, for a
given set of human-defined objectives, generate outputs such as content, predictions, recommendations,
or decisions influencing the environment they interact with”
-European Commission definition
A1.1 Introduction and Purpose

Countries have historically and continuously navigated a global technological phenomenon because of its
impact on the social and economic attributes as part of advancement in technological revolution. When
such navigation signals, it is expected that national governments must direct and plan for adoption of such
an emergence. AI technology is now such a phenomenon.
The advancement in AI as a technology has created a breeding storm and necessitates that each sovereign
country to have a deep look at its propensity, impact and pitfalls towards its citizens and needed
governance systems.
There are various definitions and descriptions of AI from the main global bodies and industries which
require synergizing and alignment. AI is described “as systems that exhibit intelligent behavior which quickly
analyse various activities and specific environments, then make independent decisions with the aim to
achieve specific [socio-economic goals] {own emphasis}” (OECD, 2018).
McKinsey and Company (2023, April 24) define AI as a “machine’s ability to perform the cognitive functions
usually associated with human minds, such as perceiving, reasoning, learning, interacting with an
environment, problem solving, and even exercising creativity”.
Boston Consulting Group (BCG) also defines AI as “a field of computer science that aims to create machines
or systems able to perform tasks that normally require human intelligence – tasks involving reasoning,
learning, decision-making, or creativity”.
Assessing the various definitions of AI at the global level, the key result from all these definitions is directed
by the effort to seek a neutral global understanding which will permit for growth for AI for all humanity.
Blumberg even remarked that “humans and machines are a match made in productivity heaven” (Blumberg,
et al, 2023).
SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 1

Historically, the human species has witnessed machine impact since the invention of the wheel that
revolutionized agriculture, it can be concluded that machines have made the life as it is, possible. It is also
attributed that the 20th century theoreticians such as Alan Turing, a computer scientist and mathematician
to envision a future where machines could perform functions faster than humans, but machines would not
replace humans in any stage of the future.
The overall purpose and objective of this document is to craft a South African National AI Plan. This plan
will inform the creation of the certainty on the AI legal and governance environment and AI regulatory
landscape that would be solicited from stakeholder inputs at a collaborative level. It will also contribute
towards the various African continent coordinated AI activities.
This plan envisages to create a better AI future use through the following:
• Creation of policy and regulatory experiments
• A set of positive goals for what South African society requires from AI.
• Management of negative AI impacts on society and industry.
• Building an understanding of the AI technological possibilities.
• Proving certainty to society on this rapidly evolving AI technology through flexibility and
accommodation of skills, software, innovations, and applications.
A1: 2 Global Discourse: Politics of AI; Economics of AI; and Militarization of AI

According to Brynjolfsson and Unger (Oct 2023) “the collective decisions we make today will determine
how AI affects productivity growth, income inequality, and industrial concentration.” They further state that
there are good reasons to take seriously the growing potential of AI systems because they exhibit intelligent
behavior and abilities on learning, reasoning, and problem solving with view to transform the economy as
opposed to technical advances of the past years.
The advancements on AI globally have created a global discourse for countries who are looking towards
AI as a tool to advance their global positions aspirations. The AI advancements provides and ability for
countries to traverse terrains that were previously difficult to navigate.
Agreements are being completed at the global levels to address strategic matters such as Sustainable
Development Goals (SDGs), COP-17 obligations, and maintenance of the leadership on the global Gross
Domestic Product (GDP) by the largest economies.

AI may affect society as a whole and the economy and other areas such as national security, politics, and
culture, however the effect on the macroeconomic terrain is on:
• Productivity growth
• Labour market, and
• Industrial concentration
According to updated data on 04 December 2023, the current global GDP is indicated that it has grown
towards the positive side of $125 trillion. The United States is leading with more than $26, 945 trillion;
followed by China at more than $17, 786 trillion; Germany by more than $4, 430 trillion; Japan by more
than $4,231 trillion; India by more than $3.730 trillion and United Kingdom (UK) by $3. 332 trillion (Forbes
India, 2023). South Africa global GDP position in 2023 is $380, 91 billion which puts it second to Nigeria
in the African continent at $477, 38 billion.
It is indicated that AI has the potential to impact global productivity growth increase positively, the impact
on labour market through creation of creation, substitution and complementary situation between humans
and machines.
Several various debates on AI at a global level touch on the realm of politics related to AI, weaponization
of AI capabilities and economics emanating from AI. Thomson Reuters Foundation AI Governance toolkit
(www.trust.org) measured the complexity of AI governance and found that there are several global
concerns at a country level which are at a different phase and need to be resolved through governance
instruments.
AI Studies on its Impact on the African Continent
According to the report titled “AI in Africa: Unlocking Potential, Igniting Progress, A Working Paper”
produced by Access Partnership (September 2023), it is elaborated that “one of the foremost challenges in
Africa’s AI adoption is the potential to exacerbate the current social and digital divides.” However, the report
further states that “AI is already being used to address a wide range of African and global humanitarian
challenges, from predicting floods and earthquakes and improving maternal health outcomes, to protecting
endangered species and safeguarding food security.”
There is an understanding in the African continent, AI can assist in improving productivity and jobs creation
outcomes across sectors like manufacturing, agriculture, financial services, and natural resources that are
key to African economic growth. There is a challenge which remains which is the availability of detailed
research which practically demonstrates how AI can impact Africa on these sectors.

The impact of AI on society research within the African continent is limited research especially when needed
to focus on several key AI-related topics as depicted on table below, thus necessitating a need of deeper
research to understand local impact and inform decision-making.
AI Governance
The Africa continent is witnessing increasing positive advocacy approaches on the future of AI governance
which outlines various governance instruments. Some of the AI governance instruments considered are
as follows:
• AI Guidelines and Standards – these are frameworks to guide AI developers, users and policy
makers on what AI software development and applications comply with the existing AI paradigm.
• AI Government Strategy – as a high-level plan or approach that outlines goals, priorities, and
actions that the government intends to achieve with AI.
• AI Policy and Law – a detailed and specific set of guidelines, rules, principles that will guide decision
making and actions in the AI area. This will be road mapping exercises linked to domestic and
international AI prescripts.
• In terms of the AI Law, there is a codified approach which will be enforceable through rules that
discourage non-compliance as it will be accompanied by penalties as evidenced by EU-AI Act.

• AI Ethical Foundations – this relates to a set of ethical and principles conditions linked to human
rights, harmful actions and protection of personal information and data.
At a continental level, particularly in Africa, there is a strong emphasis on AI supporting sustainable

development goals (SDGs). At the country level evidence is emerging where various domestic actions are
undertaken by countries such as Mauritius, Egypt, and Rwanda amongst others.
In a bigger picture approach, there has been location of relevant national government departments such
as communications, science and technology that are playing a role of AI in the areas such as healthcare,
mining, manufacturing, and agriculture.
The awareness of the pitfalls and benefits of AI is looked at the risks and biases that would sustain digital
divides, digital inequalities, and data disparity due to lack of adequate digital infrastructure currently
experienced.
AI is a serious policy concern for the African continent and the African Union (AU) has encouraged the
need to determine countries AI governance measures. There has been many activities and guiding
instruments for African countries such as the Smart Africa AI Blueprint, Sharm El Sheikh Declaration, Malabo
Convention, AU Digital Transformation Strategy, and some reports such as “AI for Africa: AI for Socio-
Economic Development amongst others.
The are still main questions around AI impact such as “How many jobs are going to be lost because of AI?
What is the best estimate? What fraction of labour force is exposed due to AI adoption, especially
Generative AI?
IMF studies report indicates that it varies country to country. The examples demonstrated is in the United
States and United Kingdom where 60-70% of labour force is exposed to AI.
In India where the large portion of workforce is in Agriculture exposure is around 30% and in Brazil and
South Africa exposure will be 40%. What can be synthesizes is that in South Africa where the total labour
force is estimated to be 24 267 566, then 40% means almost 9 707 026 job opportunities will be exposed
to AI.
AI has the potential to add and create value in the private and public sectors more efficiently. On the
gender exposure, it is remarked that women who tends to be in retail sector or services industries there
will be a greater exposure to AI.

Countries where agriculture drives economies and women largely employed at such sectors, AI will not do
much damage. The agriculture sector in South Africa is slightly less than 2,5% but Industry and Services
are 25% and 62, 61% respectively.
Role of Government in AI
Since the Dartmouth Summer Research Project on Artificial Intelligence in 1956, where the term AI was
coined by Professor John McCarty, data and AI have mainly been fields of specialists in academic
institutions, government, and private research organizations. However, recent years have been
characterized by a shift of interest towards the public sphere to include political actors, the public, and
private companies in nearly all major sectors.
AI has the potential to deliver more benefits to humanity than any new technology in the last century. This
AI benefit is also matched by great global and human potential harms.
BCG indicates that “governments need to start looking to learning from the AI leaders when it comes to AI
implementation. Many governments have begun to implement AI across various small-scale pilots. But
they are still limited to experimentation, and few have achieved true AI at scale.”
The BCG report further advises that for AI “to implemented effectively and generate benefits for public-
sector organizations in three ways: smarter policymaking, reimagined service delivery, and more efficient
operations. Thus, the technology can help governments better meet the needs of their citizens while
making better use of taxpayer” revenue for growth.
Governments around the world are facing complex challenges ranging from youth bulge, ageing
population, climate change etc. What comes across clearly is that the public sector in conjunction with the
other stakeholders such as the academic institutions and the private sector must continually evolve to meet
the demanding needs of the citizens.
McKinsey study conducted recently, found that almost 80% of public sector transformation initiatives tend
to fail or faulter. It is however, noted that South Africa is working to buck this trend.
In harnessing AI, governments must build levers that are key to driving the phenomenal absorption of AI
in the whole of society which are as follows:
• Expanding connectivity to the Internet and Digital Infrastructure
• Fostering the plans to investing in people through talent and skills development
• Enhance an ecosystem that promotes competitive and open markets

• Accelerating technological innovation and supporting the growth of next-generation technologies,
mainly AI and Cloud Computing
A1: 3 Investment in AI Technologies and Applications

According to AI Media report (2022), AI in Africa is a cross cutting technology that has the potential to
impact many industry sectors, society, and the future of work in Africa. The global AI market was valued
at $328 billion in 2021 and is projected to have grown from $387 billion in 2022 and projected to $1,394
billion by 2029.
This economic growth is exhibiting a Compounded Average Growth Rate (CAGR) of 20.1% with some
estimates suggesting it will create a $15.7 trillion contribution to the global economy by 2030. With the
projection that South Africa’s economy will be impacted on a positive economic growth of +1% to national
GDP increase. A figure below demonstrates a state of play of AI in the African continent.
Also, a figure below demonstrates the funding raised by AI Strat-ups in the African continent since 2019 –
April 2022.

The figure below demonstrates the type of AI Investments taking place in the African continent in terms of
numbers and categories. These AI investments are enabled by a need for a conducive Technology Foreign
Direct Investments (TFDIs) and the enabling Visa Regime and related environments.

A1: 4 Global Concerns on AI
Globally it is believed and understood that AI has potential to drive significant advances in so many social
and economic fields thus making connectivity more meaningful and accessible to everyone.
It is also observed that new business models and many industries are quickly shifting towards the
implementation and maturity of AI-driven products and services. These product and services are taken
up at an amazing speed by humanity and have the potential to change the world as we know it.
There are several matters globally, however that need to be addressed substantively when confronting the
emergence of AI which amongst others are:
- Data as a meaningful developmental tool.
- Digital skills and Job Impacts (Employment)
- Global Bodies Formation and Approaches on AI

- Ethical use of AI
- Funding and Investments – Protection of Intellectual Property and Patents
- AI Start-ups and New Emerging Business (Tax and VISA Regime and Maturity Assessment Models)
A1: 5 Main Types of AI Discourse

Many people in the world currently have knowingly or unknowingly interacted with AI through various
tools such as voice assistants like Siri and Alexa that are founded on AI technology and applications. Also,
in the customer service environment there are chatbots that sprung out to help navigate some URLs and
websites.
Mainly there are two types of AI making rounds in the global discourse, Applied AI (AAI) and Generative
AI (GAI).
AAI is model or a system where artificial intelligence is used or applied to real world problems or global
wicked problems and has a serious implication for governments and industries. This type of AI can help
or assist governments to grow economic and social sectors value-chains through bringing efficiency and
elimination of human inconveniences. In the industry and business levels, this AI has the potential to make
them more profitable and introducing new business models.
GAI is an AI model that generates or creates content in a response to prompt. These are AI tools such as
ChatGPT (Chat Generative Pre-trained Transformer) and DALL-E (art generated tool) amongst others
(McKinsey, 2023).
Boston Consulting Group (BCG) defines GAI as “a set of algorithms, capable of generating seemingly new,
realistic content – such as text, images, or audio – from training data.
GAI tools can produce a wide variety of credible writing in seconds, then respond to a user’s critiques to
make the writing more fit for purpose. This has implications for a broad range of industries, from IT and
software organizations that can benefit from the instantaneous code generated by AI models to
organizations in need of marketing copy. Use generative AI resolution versions of medical images, as an
example with the time and resources saved, government can pursue new business opportunities and the
chance to create more value.

A1: 6 Components of AI- AI System; Machine Learning; Neural Networks; Deep Learning
AI has largely a common definition available from most of the global recognized organizations and a
layered approach in terms of the components. It is, however, not to be complacent that everyone in the
national psychic has the same understanding. The diagram below demonstrates how the understanding
of AI and related components could be viewed.
Diagram 1: Components of AI Ecosystem
AI is a machine’s ability to perform the cognitive functions we associate with human minds, such as
perceiving, reasoning, learning, interacting with an environment, problem solving, and even exercising
creativity.
AI techniques are increasingly using massive amounts of alternative data sources and data analytics
referred to as ‘big data’. The image below states that Big Data has four (4) variations, volume, variety,
velocity, and veracity. The image below demonstrates how the four elements of Big Data are considered.

AI is a machine’s ability to perform the cognitive functions we associate with human minds, such as
perceiving, reasoning, learning, interacting with an environment, problem solving, and even exercising
creativity. AI techniques are increasingly using massive amounts of alternative data sources and data
analytics referred to as ‘big data’. The image below states that Big Data has four (4) variations, volume,
variety, velocity, and veracity. The image below demonstrates how the four elements of Big Data are
considered.
Machine learning is a form of AI based on algorithms that are trained on data. These algorithms can
detect patterns and learn how to make predictions and recommendations by processing data and
experiences, rather than by receiving explicit programming instructions. The algorithms also adapt in
response to new data and experiences to improve their efficacy over time.
The volume and complexity of data that is now being generated, too vast for humans to reasonably reckon
with, has increased the potential of machine learning, as well as the need for it. In the years since its
widespread deployment, which began in the 1970s, machine learning has had an impact in several
industries, including achievements in medical-imaging analysis and high-resolution weather forecasting.

Such data feed machine learning (ML) models which use such data to learn and improve predictability and
performance automatically through experience and data, without being programmed to do so by humans.
Some computers have now crossed the exascale threshold, meaning that they can perform as many
calculations in a single second as an individual could in 31,688,765,000 years.
But it’s not just about computation. Computers and other devices are now acquiring skills and perception
that have previously been our sole purview.
Many people have probably interacted with AI even if you didn’t realize it with voice assistants like Siri and
Alexa are founded on AI technology, as are some customer service chatbots that pop up to help you
navigate websites.
The neural network can then make determinations about the data, learn whether a determination is correct,
and use what it has learned to make determinations about new data. For example, once it “learns” what
an object looks like, it can recognize the object in a new image. Three types of artificial neural networks
used in machine learning as defined by McKinsey study:
• Feed-forward neural networks (FfNNs) are simple neural networks first proposed in 1958, information
moves in only one direction. Meaning you can feed, or input, data into the model, then “train” the
model to predict something about different data sets. They are used in banking, among other
industries, to detect fraudulent financial transactions.
• Convolutional neural networks (CNNs) are type of feed-forward neural network modeled on the
makeup of the animal visual cortex, the part of the brain that processes images well suited to
perceptual tasks, such as identify bird or plant species based on photographs. It is suited diagnosing
diseases from medical scans or detecting a company logo in social media to manage a brand’s
reputation or to identify potential joint marketing opportunities.
• Recurrent neural networks (RNNs) are a neural network whose connections include loops, meaning
the model both moves data forward and loops it backward to run again through previous layers
helpful for predicting a sentiment or an ending of a sequence, like a large sample of text, speech, or
images.
Deep learning is a type of machine learning that can process a wider range of data resources (images, for
instance, in addition to text), requires even less human intervention, and can often produce more accurate
results than traditional machine learning. Deep learning uses neural networks based on the ways neurons
interact in the human brain to ingest data and process it through multiple iterations that learn increasingly
complex features of the data.

The COVID-19 crisis has accelerated and intensified the digitalization trend that was already observed prior
to the pandemic, including around the use of AI. Global spending on AI is forecast to double over the
period 2020-24, growing from USD50 bn in 2020 to more than USD110 bn in 2024 (IDC, 2020).
It is remarked that humans and machines are a match made in productivity heaven. The human species
wouldn’t have gotten very far without our mechanized workhorses. From the wheel that revolutionized
agriculture to the screw that held together increasingly complex construction projects to the robot-enabled
assembly lines of today, machines have made life as we know it possible.
Despite their seemingly endless utility, humans have long feared machines on the possibility that machines
might someday acquire human intelligence and strike out on their own.
A1. 7 Regional vs National AI Approaches and Frameworks

There is 70% plus global internet adoption in 2022 and more than 60% online mobile usage across major
markets (AI Media Group, 2022). Generative AI and ChatGPT in particular, swept into the world’s collective
consciousness and made us all sit up and take notice The application reached 100 million monthly active
users just two months after launch and swiftly earned the honour of being the fastest-growing consumer
application in history (Ekow Duker, co-Founder of Zindi Africa).
According to McKinsey research the real value for the Middle East’s Gulf Cooperation Council (GCC)
countries could be $150 billion or equivalent to 9 percent or more of those countries combined Gross
Domestic Product (GDP).

The figure above as researched by Access Partnership shows that the Sub-Saharan Africa could gain up to
$136 billion worth of economic benefits by comparing the AI sector AI impact in South Africa, Ghana, Kenya
and Nigeria. As previously reported that depending on each country key economic sector contributor, AI
will determine the value derivative. For example, Nigeria key sectors stands to benefit from AI at a greater
scale compared to South Africa which its key sectors will be somehow negatively impact to a greater extent.
However, all these countries will enjoy greater benefits from the adoption of AI.
In adopting AI at a national stage is by creating and understanding and acknowledging of the global digital
transformation changes happening at the United Nations (UN) levels and many global platforms.
Secretary-General (SG) of the UN report on the preparation for the Summit of the Future in September
2024, indicated that a Digital Global Compact (DGC) must be developed encapsulating AI is an embedded
technology. The DGC will be driven by focus on Digital Transformation and Climate Change as key drivers
for global changes with AI as the center of gravity.
A1: 8 South Africa’s AI Conceptual Framework and Approach
South Africa (during its African Union Presidency) was instrumental in contributing to a pan-African “AI for
Africa Blueprint” part of the SMART AFRICA initiative supported by the German Development Cooperation
(GIZ) and the Smart Africa Secretariat. The AI for Africa Blueprint helps member-states towards developing
policies, strategies and plans that ensure growth and prosperity within the context of 4IR digital revolution.
In November 202, the Department of Communications and Digital Technologies (DCDT) launched an
Artificial Intelligence Institute of South Africa (AIISA) [www.aii-sa.co.za] to encourage the take-up of AI and
localised AI solution.
In addition, the Department of Higher Education, Training, and Innovation (DHET&I) established the WEF
Affiliated AI Centre which focuses on AI Ethics and the Centre for Artificial Intelligence Research (CAIR)
which focuses on AI Research.
These current approaches were informed and directed by the several prescripts but mainly from the
Presidential Commission on the 4th Industrial Revolution (PC4IR) report. It was adopted at Cabinet meeting
on August 26, 2021. The establishment of the institute sums up the actions that South Africa’s government
will take to advance the take-up of AI in both private and public sector, to increase the relevant skills and
research.

The envisaged South African National AI plan will rely on the following pillars:
• Boosting AI in the government, economy, digital skills through research and development. and the
legal environment.
• Government-as-Platform approach to boost uptake of AI in both public sector and wider economy.
• AI public skilling courses and raising AI awareness.
• Creation of AI sandboxes for testing AI localised applications.
• Collaboration with the private sector, academia, and civil society to identify AI opportunities to
advance society.
• Develop designated AI local innovation and development of Start-up grants for developing machine
learning based solutions.
• There is a need for a separate legislation regime that will demonstrate the importance of AI as a field
of technology that is becoming increasingly popular.
• AI has the potential to deliver real value in South Africa.
Section B: AI – South African Landscape, Opportunities and Pitfalls
B1. AI Understanding within South Africa’s Context
The greatest attention currently on AI is because it is now able to perform the functions that were
exclusively human brain function. AI is a machine’s ability to perform the cognitive functions we usually
associate with human minds. AI systems are machine-based systems with varying levels of autonomy that
can, for a given set of human-defined objectives, make predictions, recommendations, or decisions.
The image below demonstrates how the AI system through data accumulation and interpretation can
perform brain functions as used by many others.

In the context of South Africa where there is a challenge with job security and high unemployment, the
use of AI should be dovetailed with the need to improve such situations. It is remarked that a mere
adoption of AI within the administration work phase could save South Africa $9 billion an annum through
improvement of absenteeism.
One of the significant aspects about AI as a general-purpose technology is that is not a standardized
blanket approach technology. It offers the user or adopter the opportunity to tailor make it to address the
specific challenges at that moment and in future. For example, there are countries that looked at AI to
help solve their national problem by addressing the national question which was based on solving the
aging population crisis.
This was applied with the understanding that AI can provide the ability to analyze, diagnose, predict, and
automate a problematic situation at a faster speed.
AI is revolutionizing human life and unlocking a whole new world of innovation. Its prevalence will affect
nearly every society aspect. The AI Industry and developers are required to demonstrate responsible
behavior and ethical use.
AI solutions and applications must consider that they are designed to ensure safety, reliability, and ethical
practices. There will be a need to create trustworthy regulatory principles and governance systems that
will guarantee levels of accountability, inclusivity, and transparency that will promote human life and prevent
harm.
B1. 1 Adoption of AI: SA AI Approach PC4IR

In South Africa’s situation, the PC4IR report has outlined 10 national features that need to be understood
before deciding on AI approach. These features include but not limited to youth bulge, gender inequality,
internal migration, income disparities, human development index, skewed economic structure, regression
in energy and water supply, domestic workers dominance, weak economic growth, high corruption,
reduced global competitiveness, amongst others.

The PC4IR also outlines 8 key dimensions that should guide the development of an AI national plan
approach which are:
• service delivery,
• preparation through experimentation,
• private sector capabilities,
• technological clarity and commitment,
• government as an organizing player,
• human capital development, and
• regulation and ethics.
There is also un understanding provided by the PC4IR report when it comes to policy development and
formulation that outlines required approach: policy must be inclusive in its process, integrated
infrastructure framework, financial incentives, adaptiveness, and socio-economic impact assessment.
Companies and organizations in South Africa have started with the implementation of both the applied
and generative AI technologies in areas such as chatbots, job seeking tools, quantum analysis and data
traffic management.
B1. 2 Applied and Generative AI: Impact in South Africa

There is a need to understand the various forms of AI especially in two (2) formats, Applied AI, and
Generative AI. Applied AI simply, artificial intelligence applied to real-world problems has serious
implications for the business world. In this aspect, the value of AI is not in the systems themselves but in
how companies use those systems to assist humans and their ability to explain to shareholders and the
public what those systems do in a way that builds and earns trust.
The PC4IR has placed emphasis on the adoption of applied AI in advancing the economic growth prospects
of South Africa. The key economic sectors such as manufacturing, transportation or network industries
require an applied AI to function optimally.
Generative AI is an AI model that generates content in response to a prompt. AI tools like ChatGPT and
DALL-E (a tool for making AI-generated art) have the potential to change how a range of jobs are
performed.

The full scope of that impact, though, is still unknown, as are the risks. Questions that can be answered
are, how generative-AI models are built, what kinds of problems they are best suited to solve, and how
they fit into the broader category of AI and machine learning.
Generative-AI tools can produce a wide variety of credible writing in seconds, then respond to a user’s
critiques to make the writing more fit for purpose. This has implications for a broad range of industries,
from IT and software organizations that can benefit from the instantaneous code generated by AI models
to organizations in need of marketing copy. Use generative AI resolution versions of medical images, as
an example with the time and resources saved, government can pursue new business opportunities and
the chance to create more value.
Generative AI will have heavy impact on the educational, art and culture, and research and development
areas. There is a need to shape up and build defensive all the address possibility of fake and misleading
facts through proper AI regulation mechanisms.
B1: 3 State Considerations in terms of Law, Governance and Regulation

The deployment of AI in major sectors is expected to increasingly drive competitive advantages through
two main avenues:
• Cost reduction, and
• Productivity enhancement.
Policy makers and regulators have a role in ensuring that the use of AI is consistent with regulatory aims
of promoting stability, protecting consumers, promoting integrity and competition. Policy makers should
consider supporting AI innovation in the different sectors.
Emerging risks from the deployment of AI techniques need to be identified and mitigated to support and
promote the use of responsible AI. Existing regulatory and supervisory requirements may need to be
clarified and sometimes adjusted, as appropriate, to address some of the perceived incompatibilities of
existing arrangements with AI applications.
The application of regulatory and supervisory requirements on AI techniques could be looked at under a
contextual and proportional framework encouraging the use of AI without unnecessarily stifling innovation.
Policy makers should consider sharpening their focus on better data governance. Policy makers should
consider disclosure requirements around the use of AI techniques in the provision services and that it may
impact outcome.

Regulators should consider how to overcome the perceived incompatibility of the lack of explainable in AI
with existing laws and regulations.
Policy makers should consider requiring clear AI model governance frameworks and attribution of
accountability to help build trust in AI-driven systems. Explicit governance frameworks that designate clear
lines of responsibility for the development and overseeing of AI-based systems throughout their lifecycle,
from development to deployment, could be put in place. Internal model governance frameworks could
be adjusted to better capture risks emerging from the use of AI.
What is needed to be in place to Achieve with adoption of AI in South Africa?
The era of Digital Transformation provides states and nations options to navigate. These options are related
to achieving phenomenal economic and social growth which will yield returns that are realizable. The
options that are readily applied within the digital transformation are:
• Digital Economy
• Gig Economy
• Circular Economy
A choice or combination of these options have the requirements and yields for when a country or nation
decides on a particular choice within the digital transformation can provide a springboard. The table below
shows a snapshot of what each choice has enabling requirements and outcomes that will be realized. It
should be borne in mind that should a specific option be made further unpacking will become a driving
policy and strategies that would be compounded by needed regulatory framework.
Choices/Options Enabling Requirements Outcome
Digital Economy Robust 80- 100% Internet Disruptive Development: Remodeled

Connectivity Access by society Systems that will provide Global
Leader and Enhanced
Competitiveness
Innovation/Gig Economy 80-100% Funding on Game Changer Development:

Invention and Local Innovation with global reach
Innovation.

Choices/Options Enabling Requirements Outcome
Large segment of society

must be of young age with
government having a lost risk
appetite.
Circular Economy 80-100% consumption based Sustainable Development: Recyclable

on net importer of goods and Greening Systems
In explaining the table above, will be guided by the understanding of how each option is defined within
the context of a particular nation. Defining the Digital Economy, a unified definition is that “is a worldwide
network of economic activities, commercial transactions and professional interactions that are enabled by
information and communications technologies (ICT), meaning an economy based on digital technologies”
(www.techtarget.com).
Digital Economy require that a nation or society must have 80-100% must be connected to the Internet,
meaning that government, industry, academia, and civil society must use Internet in most it’s their activities.
The outcome is that the current systems must be remodeled, and change must be deep.
Innovation/gig economy is defined as “a field of economic theory and applied/experimental economics that
emphasizes innovation and entrepreneurship” (www.valuer.ai). The choice of Gig Economy requires that
the society must be driven by youth bulge where at least 75% of the population is younger than 40 years
old. In this option, the government must be willing to fund local inventions and innovations with the
intention to have a game changing innovation that will be globally adopted.
Circular economy is defined as “a model of production and consumption, which involves sharing, leasing,
reusing, repairing, refurbishing and recycling existing materials and products as long as possible”
(ww.europarl.europa.eu). The choice of a Circular economy relies on the society which does not actually
produce goods at a large scale but relies on importing goods from other nations. This means that whatever
is imported must be made longer term longer and recyclable for longer-term.
South Africa is in the middle strata of these economic models because they possess snippets of elements
of each and not through the determined thresholds. The acquisition of AI technology determines the
choice of an economic model pursuant to social and economic growth. South Africa has made a choice
to pursue a digital economy model as the digital transformation agenda, thus its use and adoption of AI
will be demonstrated as such.

Frameworks for appropriate training, retraining and rigorous testing of AI models could be introduced
and/or reinforced to ensure that ML model-based decision-making is operating as intended and in
compliance with applicable rules and regulations.
Regulators should consider promoting the ongoing monitoring and validation of AI models, which are
fundamental for their risk, as one of the most effective ways to reinforce model resilience, prevent, and
address model drifts.
Appropriate emphasis could be placed on human primacy in decision making when it comes to higher-
value use-cases. Policy makers could consider the increased technical complexity of AI, and whether
resources will need to be deployed to keep pace with advances in technology. The role of policy makers
is important in supporting innovation in the AI sector and protecting the citizens from consuming AI
products that are harmful to their privacy, security, and dignity.
B1: 4 South Africa’s Aspirations: Global and Continental
South Africa’s global AI Aspirations are guided by the following objectives:
• To promote national character as a global shaper in the domain of AI.
• AI Skills Aspiration - South Africa aims to enhance SA human capital with a steady local supply of
AI-empowered talents to cope up with the disruption that will be caused by this technology.
• AI Policy And Regulation Aspiration – South Africa will be part of legislative national transformation
of the African continent in enacting the most welcoming legislation AI businesses and talents.
• AI Investment Aspiration – South African government will support increase assets and enable the
economy by attracting local and foreign investors to identify and fund AI investment opportunities
in SA.
• Research and innovation initiatives will target top human capital talents to nurture their skills through
state-of-the-art R&I environment in new/smart cities to spearhead innovation and impact creation.
• Ecosystem Initiatives aim to stimulate Data & AI adoption across giga projects and across major
cities, improving productivity, quality of services and wellbeing

Section C: South African AI Plan
C1 Priority Focus Areas and Enablers

A national AI Plan Mapping Steps
2023 2024 2025 2026
With foresight and a commitment to act, South Africa can capture the opportunity offered by technology,
manage the AI risks, and ensure that the gains are broadly shared. There are four core priorities for a bold
AI national plan of action that is co-owned by government, business, labour and educational leaders.
Provide the state with predictive maintenance, diagnostic ability, analytical models for policy and regulatory
systems, and automated service delivery channels.
In terms of AI as a technology tool of choice and the impact it must have on the social and economic
sectors is to achieve the following outcomes:
• AI Predictive maintenance abilities.
• AI Logistics optimization and Automated services
• AI Diagnostic abilities
• AI Analytical abilities
AI National Major Initiatives Required to Scale-UP
• Launching and events  AI Insititions
• Operationalisation and Activiting
• Building AI Data Achitecture

• AI delibarate investment channels
First, South Africa can shape a smart approach to embrace AI for the benefit of all citizens. The plan can
help define the country’s AI agenda, including prioritizing AI industries where South Africa has or can
develop a competitive edge in technology-driven solutions, such as mining and manufacturing. The
envisaged national plan AI of action can also scope and size the risks that AI adoption poses to jobs and
the economy; it can also assist develop the necessary AI national plan.
Second, South Africa can take decisive steps to invest in AI skills generation plan. The plan could help
rethink the country’s education and learning ecosystem to foster broad participation in the AI opportunities
of the future. It can create transparency on the AI skills that the economy will need in future, and help
educational institutions plan to meet that need. It can trigger a review of South Africa’s basic education
system, to strengthen the focus on teaching quality, literacy, and STEM subjects - science, technology,
engineering, and mathematics. It can promote a stronger technical and vocational education and training
(TVET) system to ramp up the development of AI technical skills. And it can help drive improvement in the
graduate conversion rate of South Africa’s education system.
Third, the AI national plan can set out a national strategy to mitigate the AI impact of adoption. That would
include identifying jobs that will not be portable and identifying ways to reintegrate displaced workers into
the AI workforce. For example, South Africa can increase investment in AI to boost employment
opportunities. That can include providing focused support to AI start-ups that are producing AI products
for export or import substitution in sectors such as advanced manufacturing and agro-processing, and
which have prospects for rapid growth. South Africa can also develop support for individuals moving from
formal employment into the ‘gig’ or “digital” or “circular” economy. Finally, the plan can propose ways to
rebalance AI social-support funds and mechanisms to support those who cannot be AI integrated into the
workforce.
Fourth, the AI plan can help foster a step-up scenario, where AI job creation increases over and above a
trendline scenario because of societal and policy choices enabled by increased productivity and economic
growth. Closing the AI infrastructure and needed gap will require an increased levels of investment in the
South African AI sector economy and could be a major as this could be a source of labour demand in
years to come.
A further opportunity is embracing the transition to clean energy through AI, including increasing the share
of renewables and making buildings, transportation, and technologies more energy efficient.

AI Dimensions AI Components AI Activities
Skills Data and Literacy • Instil basic data and AI skills for the
national workforce and students.
Data and AI Specialists
• Build sustainable pool of national
Data and AI Experts specialization in data and AI.
• Nurture data and AI scientists

through local and international
partnerships at an academic and
professional levels.
Policy and Regulations Activate policy and regulatory • Data generation AI generative and
frameworks. applications.
• Attract local and foreign

companies to AI sectors through
Companies’ incentives
incentives programmes.
programmes.
• Attract foreign talents data and AI
sector.
Talent’s incentives
programmes
AI Investment AI Investment Funds • Stimulate data and AI investments

through targeted investment
funds.
Investor support programme

• Activate specialised programme to
support investors.
Administrative services

Research and Innovation R&D innovation excellence. • Elevate data and AI R&D
innovation capabilities in
National innovation test beds.
universities and research centers.
• Create an attractive data and AI

test bed environment in mega-
projects and smart cities.
AI Ecosystem Data and AI Sector strategies • Develop data and AI domain in

sectors identified as priority
Data and AI Platforms
strategies.
Smart cities • Accelerate data and AI technology

adoption through accessible
Government Systems
digital platforms.
• Drive the smart city agenda in

towns and cities.
• Adopt data management practices

and use of AI in government
organisations
C1. 1 Discussion on Enabling Environment

South Africa is observing three (3) trends in which the adoption of AI is possible:
• Firstly, technologically, the data from machine learning is tied to the wide availability of large amounts
of data that could be used through cheap computing power.
• Secondly, economically, progress on adopting AI is based on significant investment input and

advancement of venture capital in technological companies and start-ups.
• Thirdly, culturally, the digital content generation drives the digital transformation global agenda as it
enables AI platform generation.
Which South African Economic Sectors can benefit from machine learning machine and deep learning?

The image above demonstrates the plethora of sectors that have been advanced in terms of structure and
size in terms of maturity, semi-maturity, and least maturity.
The McKinsey study collated more than 400 use cases of machine and deep learning across 19 industries
and nine business functions. Nearly all industries can benefit from machines and deep learning.
The identification of key economic sectors that could benefit from application or adoption of Applied and
Generative AI should be harnessed and supported through the lens of acknowledging the changing
business models.
EDUCATION: Integrating Data and AI in education to align the education system with labor
market needs and improve the student’s journey
GOVERNMENT: Assimilating Data and AI in Government to create a smarter and more efficient
public sector
HEALTHCARE: Integrating Data and AI into healthcare to increase access, enhance

preventative care, and accommodate growing demand

ENERGY: Integrating Data and AI into Energy in increase capacity, enhance efficiency, and
develop adjacent industries
MOBILITY: Incorporating Data and AI in Mobility to build a regional hub, create smart city
Mobility technology, and enhance traffic safety in cities
APPLYING AI in terms of Process Steps
The steps that are outlined in the above image demonstrates the number of inputs and planning required
to adopt by the various sectors in creating a value-based design that would guide the narration of AI in
SA planning process.
The number of AI Institutes and Centres within SA must endeavor to share a national design data and
models that would facilitate the quick adoption of AI within the SA national economic and social footprint.
AI National Targets in Catapulting and Dovetailing NDP 2030
• 30% of the total workforce trained on basic AI data management and AI literacy skills.
• 20 000 local AI specialists in workforce.
• 5 000 AI experts
• Top 20 of countries in Applied index

• Exemplary AI legislative aspects and regulation
• 100 - 300 AI startups
• 70 Bn Government, local and Foreign Direct investment AI
• Top 20 countries in peer reviewed SA AI research and development publications
Why AI Adoption matters in SA?
South Africa’s position regarding the stack choices mentioned above seems to sit in the middle of the
choices. The ambition to make the South Africa to be at the place where AI is made reality goes in parallel
with the national priorities under NDP 2030.
The country is going through an unprecedented transformation, and we are aiming to maximize the
benefits provided by AI to support and accelerate this transition. South Africa needs to ensure the
continuity of this endeavour throughout the design and implementation of the different phases of the
policy and strategy by focusing on six broad dimensions. The dimensions are policy and regulations
imperatives, generative skills, needed investment, research and innovation and the data and AI ecosystem.
These dimensions will support the proposed four (4) phases related to planning for – strategy formulation,
activate initiatives, expand execution, and accelerate execution.
Under the 2030 National Development Plan (NDP) vision of the Republic of South Africa (RSA), the
Presidential Commission on the 4th Industrial Revolution (PC4IR) report and strategic framework developed
in 2021, Artificial Intelligence (AI) became a key topic to be developed and integrated in the new digital
transformation agenda
and adoption of 4IR technologies. South Africa vision on AI is defined by the intention to establish a
national AI Institute that will drive the economic and social policies designed to advance national interest
and global competitiveness.
An AI Framework was developed through the formulation of a Strategic Implementation Programme for
the PC4IR report. Below image is the depiction of the components of the AI framework.

As announced on 30 November 2022, the South African government through the Department of
Communications and Digital Technologies (DCDT) has been actively driving digital innovation by
establishing an Artificial Intelligence (AI) Institute in partnership with the academic institutions which will be
the Centers of Excellence and Hubs for the application of AI within the sector economies.
The DCDT in partnership with University of Johannesburg (UJ) and Tshwane University of Technology (TUT)
embarked on the establishment of an AI Institute was the catalysts for ushering South Africa into the digital
frontier. There are keyways that must be crafted and driven to ensure the practical success of this initiative
through ensuring programme approach, championing, funding, and sustainability activities.
Since the past years, the situation across the world economy and society has slowed down due to the
current pandemic, AI in its applied and generative format provides a key opportunity for South Africa and
the globe to test its potential.
From South Africa’s perspective, AI has the potential to close the growing digital gap that exists between
developed and developing nations, which has been made more prominent during this unprecedented
period.
This vision was part of the discussion about the launch and establishment of AIISA. This requires all of us
to think and work hard to take advantage of artificial intelligence and unleash its full potential to advance
our societies and economies.

With foresight and a commitment to act, South Africa can capture the opportunity offered by technology,
manage the risks, and ensure that the gains are broadly shared.
Four core priorities are suggested for a bold national plan of action that is co-owned by government,
business, labour and educational leaders.
First, South Africa can shape a smart approach to embrace digitisation for the benefit of all citizens. The
plan can help define the country’s digital agenda, including prioritising industries where South Africa has
or can develop a competitive edge in technology-driven solutions, such as mining and manufacturing. The
national plan of action can also scope and size the risks that digital automation poses to jobs and the
economy; as we discuss below, it can also help develop the necessary mitigation plans.
Second, South Africa can take decisive steps to invest in human capital. The plan we suggest could help
rethink the country’s education and learning ecosystem to foster broad participation in the work
opportunities of the future. It can create transparency on the skills that the economy will need in future,
and help educational institutions plan to meet that need. It can trigger a review of South Africa’s basic
education system, to strengthen the focus on teaching quality, literacy, and STEM subjects - science,
technology, engineering, and mathematics. It can promote a stronger technical and vocational education
and training (TVET) system to ramp up the development of technical skills. And it can help drive
improvement in the graduate conversion rate of South Africa’s education system.
Just as important in strengthening South Africa’s human capital, the plan can help guide both public and
private-sector organisations to create ‘retraining engines’ to enable lifelong learning. That will entail several
fundamental shifts in the country’s current approach to education and training. For example, we might
need to shift from a focus on once-off degrees to discrete certifications in specific skills.
We might develop a new taxonomy of certified skills to enable workers to ‘brand’ themselves for specific
capabilities. And, to help training providers rise to the challenge of large-scale reskilling, the plan could
define a clear accreditation framework and recommend new quality-assurance processes that focus on
application of skills rather than theoretical understanding.
Third, the plan can set out a national strategy to mitigate the jobs impact of automation. That would include
identifying jobs that will not be portable, and identifying ways to reintegrate displaced workers into the
workforce. For example, South Africa can increase investment in SMEs to boost employment opportunities.
That can include providing focused support to SMEs that are producing products for export or import
substitution in sectors such as advanced manufacturing and agro-processing, and which have prospects
for rapid growth. South Africa can also develop support for individuals moving from formal employment

into the ‘gig’ economy. Finally, the plan can propose ways to rebalance social-support funds and
mechanisms to support displaced workers who cannot be reintegrated into the workforce.
Fourth, the plan can help foster a step-up scenario, where job creation increases over and above a
trendline scenario because of societal and policy choices enabled by increased productivity and economic
growth. For example, infrastructure systems have not kept up with demand and housing shortages persist
in many countries – including South Africa. Closing this infrastructure and housing need gap will need
increased levels of investment in the South African economy and could be a major source of labour
demand in years to come. A further opportunity is embracing the transition to clean energy, including
increasing the share of renewables and making buildings, transportation, a
The established AIISA will be addressing several critical AI-related four main areas for consideration and
direction and themes.
• Shaping a New Normal through AI
• AI and Government role
• AI Governance and Regulation, and
• The Future of AI.
C1. 2 Focus Areas: What must be in place for AIISA to be Effective and Realistic
Right Data Architecture
The journey towards digital transformation is the most pressing challenge within the public sphere is
whether the infrastructure that exists is ready for innovation of tomorrow. In simple terms, the public-
sector and its partners need to ask the question: Can their existing technology architecture and systems
effectively store and manage the amount of data volumes that require the effective application of AI
systems?
Management and ownership of Data in particular, has become an essential developmental resource for
many countries. It is even remarked that is now “the new oil,”. This illustrates the possibility of immense
opportunities it can create in terms of efficiency, development of AI solutions, and public services. Major
ccountries are learning to maximize data collection opportunities, processing, and integration into
developmental agenda to provide valuable solutions. It is expected that it will gain significance, in the
short and long term. SA, therefore, has a critical role in developing the Data and AI agenda to position
the country in the league of data-driven economies.

It is reported that by 2025, about 463 exabytes of data would need to be generated globally each day and
managing this paradigm is a challenge that must be addressed imminently. The MIT Technology Review
report indicated that despite the great enthusiasm for Artificial Intelligence (AI), it was found that 77% of
organizations in the
Asia-Pacific and Japan find it difficult to digest, analyze and interpret data. This is the major task the
Artificial Intelligence Institute of South Africa (AIISA) must face head on and address as the first task.
Effective data analytics first requires a scalable and reliable infrastructure of which the sector government
departments must understand through the need to lay out public data architecture systems. It is estimated
that at least R70 billion at minimum should
be dedicated towards 5G innovation and more than R100 billion would need to be set aside to develop AI
systems and meet national cyber security needs. The public sector will have to upgrade and rebuild legacy
IT systems to support the seamless storage, retrieval, and analysis of data.
AI Adoption Programme Approach
The management and operations of the AI Institute would require that it be implemented at a programme
level within the DCDT structure and systems. This would allow for robust business continuity capabilities
which are able to reach the higher echelons of government and executive suites of the private sector.
South Africa citizens are incrementally and increasingly adopting a digital lifestyle, which means
government services are expected to be always online. The speed and agility would not sustain the long
downtime as siloed and legacy IT systems and mundane maintenance and backup processes may be
catastrophic consequences. What is not to be forgotten is the constant cyber-attacks that always threaten
the delivery of e-services.
The rationale for the programme level can be elucidated by an example where an essential service such
as a national water distribution system which requires constant monitoring of water pressure and quality
distribution is key. The AI Institute would require advice and intervene by identifying leaks and possible
contaminations as an early warning through real-time data through reduction of human intervention.
Operating at the programme level is thus critical to the smooth operations of government services, thereby
increasing timely and trustworthy interventions.
The AI Hubs must be independent on their operations and project management guided by the policy and
strategy programme of the DCDT.

The AI Hubs will develop a report mechanism will will be provided to the DCDT at quarterly intervals to
allow for proper reporting within government structures.
AI Engagement Strategy and Approach
The purpose of an AI Institute is to assist sector departments at various government levels to effectively
use AI as digital tools that will advance the value-chains and mandates. Various national, provincial, and
local governments have indicated the need for accessing and using digital technologies such as AI towards
solving their challenges.
The AI Hubs have identified catalytical projects related to the sector solutions such as transport, mining,
farming amongst others. These projects require an ability to provide proof of solution and the ability to
hit the ground running to those sector clients.
There is a need to develop a planned engagement strategy towards the identified stakeholders through
the AI Institute. The AI Institute should be the first point of call and set up the required engagement with
government departments. Once the groundwork has been laid, then the Hubs will take over the projects
on both the operational and technical elements.
The approach must be based on a solid foundation wherein a sound proposal by the AI Hub on the
identified solution must be developed as an engagement tool.
AI Advisory Council or AI Panel of Experts
There was a discussion about the need to establish an AI Expert Advisory Council which was facilitated by
DCDT. The processing of the establishment has faced some constraints internally that would have to be
resolved soon.
The AI Hubs must also determine how they will have advisory council linked to their areas of focus.
Funding, Investment and Sustainability
The adoption of AI is a general – purpose technology (GPT) will require funding related to aspects such
setting data architecture, equipment, lab tools etc. There are various
ways for funding the AI Institute and its Hubs. There is a need to ensure that a seed fund plus the launch
costs must be budgeted to ensure awareness and visibility.
The AI Institute sustainability is paramount to its impact and sustainability because it will have the whole of
society participation. Investments must be sourced or accepted from various sources which are

government, academia, private sector, and international donors. These investments would come from the
influential quarters depending on what area of AI their focus will be based.
The AI Institute and the Hubs must establish a framework to address such investments and a vehicle
mechanism to manage such investments. The key issue is that for any
sort of investments to flow, there is a need for seed funds that would propel the initial work that must be
undertaken by the AI Institute.
Structuring and Resources (Governance)
The setting-up of an AI approach by the DCDT is based on having a programme approach wherein to
avoid a brick-and-mortar physical structure outside the mandate department because AI is a national
imperative. The AI Institute role is to be the coordinator, facilitator, and policy setting organization rather
than direct operational implementation. The AI Institute would be made up of a network of AI Centers of
Excellence or Hubs with a focus on different but interlocutory areas of execution. A decision matrix must
be put in place which will allow for consistent use of resources.
The structure of the AI Institute must be modelled around the role of the networked AI Hubs or Centers of
Excellence. It is envisaged that the governance structure must be codified and agreed by all to allow for
plugging of other initiatives.
The role of the private sector involvement must also be incorporated within the government structure. The
decision-making matrix must also deal with the distinction between the management, operational and
technical aspects.

The last governance structure agreed to is demonstrated below:
During the preparation of the AI Institute establishment and launch of the UJ Hub there were various
working committees that were created which can be reviewed to ensure proper governance and required
resources. The highest decision structure was the
AIISA Launch Preparation Committee which includes the DVCs and the Acting Director-General of the
DCDT.
Focus Areas of AI Hubs
The UJ and TUT AI Hubs presented catalytical projects aligned with the focus areas which are distinct and
overlapping. There will be other AI Centers that would be added within the network already created and
their focus areas would have to be decided upon regarding how they get infused within the existing focus
areas.

There might be a need for the AI Hubs to have a workshop to align the areas of focus and build the
required engagement plans to the identified sector departments.
Benchmarking and Fact-finding Missions
The exponential rise and global focus on AI as a technology has witnessed a plethora of establishing AI
Institute or centers by countries and global organisations. There are also various private sector
developments and AI platforms being established around the world that have shown interest in the work
of the AIISA.
The AI initiatives that have approached DCDT are:
• Saudi Data and Artificial Intelligence Authority (SDAAI) – United Kingdom of Saudi Arabia
• The AI Institute of the Netherlands – Republic of Netherlands
• National institute for Transforming India (NITI Aayog) – Republic of India
• UNESCO AI Ethics Centre
• AI- Sweden
• Artificial Intelligence Technology Centre – Republic of Egypt
There is a need to conduct fact-finding missions and benchmarking studies by the AIISA to build global
cooperation and develop best practice system which will allow AIISA to be competitive and in line with
current affairs.
C1. 3 Data, AI and IOT

CORE MANDATE: DCDT’s key role is to Drive and Own the National Data and AI agenda, through the
following objectives.
1) Set the overall strategic direction of AI and Data - Define the nation- wide strategic direction of AI
and Data
2) Guide the execution of the AI strategy with government and Entities - Support and guide different
entities in strategy execution, without hindering their current Data and AI plans. Address the VISA
Regime and Global Investments access path.

3) Execute AIISA led initiative - Achieve and fulfil AIISA initiatives such as AI data consolidation and
regulation
4) Fund AI Local Initiatives, communication, and public awareness – Strategic Funding for Local AI
Innovations
C2 AI Governance and Institutional Mechanisms

C2: 1 South Africa’s AI Ethical Considerations
There are future potential AI harms will require the adoption of a set of human and ethical considerations
that have been postulated but where it is less clear what the optimal form of regulation is, such as:
• Anti-competitive behavior should the provision of key AI models, or other enablers, be concentrated
in a small number of players.
• Risks in robotic or autonomous devices that use AI to decide on actions, such as AI-enabled
autonomous cars.
• Aggressive loss of employment leads to social risks.
• Criminal behavior or other highly dangerous outcomes.
• Existential risks should AI “get out of control” and pursue goals detrimental to humanity.

• Risks that AI designed for military purposes “escapes” into the wider world and leads to enhanced
criminal behavior or other highly dangerous outcomes. (Access Partnership, 2023)
South Africa is witnessing global numerous calls on the AI ethical dimensions such as
• OECD – AI Trustworthy and Reliable use
• UNESCO – AI Ethical use
• ITU – AI for Good
• UN – AI for Development
South Africa has to navigate within these global proposals and find a solution that is aligned with the
developmental state aspirations. The cornerstone of South Africa human rights is located within the
Chapter 2 of the Constitution and the preamble. Below images demonstrate some of the approaches that
would guide the consideration of on aspects that should be considered when addressing the AI impact
and pitfalls.

C2: 2 SOUTH AFRICA AI REGULATION APPROACH
According to Access Partnership AI Policy Lab (2023) [www.accesspartnership.com], “regulators, policy
bodies, and companies have responded to this clear need with a flurry of proposals to either regulate AI or
build the evidence base to do so. However, the quality of these proposals varies considerably. Others,
meanwhile, are considering alternative approaches, such as voluntary ethical frameworks.”
Recently the most concrete regulatory prescript was from the European Commission through a EU’s AI Act.
(see below image) This has been through many years in development and is now approved by the EU
structures which is detailed over 100 pages document and includes enforcement mechanisms and more.
It is regarded as the first comprehensive piece of legislation in the global AI space and became the de
facto global AI regulation.

The observation is that other countries have varying degrees of regulation and frameworks, while it is not
clear that whether South Africa and others will follow the same approach or would come up with anything
materially better based on the environmental conditions.
Any regulatory approach to AI, in terms of South Africa, it should be guided by the following outlined
factors as per Access Partnership AI Lab:
• Global practices since AI by its nature cannot be constrained to any country.
• Responsive to Economic and Social Aspirations because of the need to ensure an all-inclusive
approach to ensure that the whole society benefit as the AI harms become apparent in many
instances.
• Built on principles, or ethics, and human rights protection that provide guidance for those working
on AI and enable even some degree of self-regulation.
There are other regulatory concerns that should be borne in mind are as follows:
• Disinformation and fake news generated through Generative AI applications and software.
• Bias and discrimination that could be exhibited by AI developers and users.
• Copyright abuse in alignment with World Intellectual Property Organisation (WIPO)
• Privacy to protect personal, private, and public data.
C2.3 SA AI and DATA Governance Structure

AI is still quite a novel technology which may not have an immediate economic impact within organisations,
as it is still being developed. However, the risk of waiting for the technology to mature means lagging
cannot be tolerated. The support required from government is for the application of governance of AI
and mitigation of the risks associated with its implementation towards significant economic impact in the
future.
The implementation of AI governance on applications is a high priority on discussion topics in continental

and global cooperation groups of the African Union (AU), SADC, BRICS+ and G20 member-states.
The relationship between Data and Cloud, IOT and AI should inform the required governance structure at
the programme and executive level. The governance structural depiction is informed by the tripartite
relations of data, IOT and AI as demonstrated below:

IMC on Digital
Cabinet
Transforma�on
Minister
AI
Management
Office
Min AI Expert
Advisory Council
AIISA
Programme
Data and Cloud

Office
AI Hubs
IOT & e-Gov
office
The governance structure and model of AI from the global perspective is guided by relationship cultivated
the role of government, private sector and civil society.
Cabinet Role – Approve and guide the execution of the National AI Plan with government entities.
Minister and AI Expert Advisory Council Role - Set the overall strategic direction of Data and AI and define
the nation-wide strategic direction.
AIISA Programme - Facilitate the execution of plan and strategic initiatives. Achieve and fulfil AIISA
initiatives such as data consolidation and regulation
AI Hubs – Support and guide different entities in strategy execution by generating AI localised solutions,
applications, skills, and software development, without hindering their current Data and AI plans.

AI Management Office – exists to motivate AI innovation and build capabilities.
Data and Cloud Office – regulate national data and drive compliance.
IOT and e-Government Office - manage national data and draw insights through data processing.
C3 AI Plan Adoption and Timelines Processing

South Africa’s Action Roadmap and Plan for the implementation of AI
The main presume is that a shift from analogue thinking towards digital thinking requires a redesign of the
systems and mindsets from the population at large.
The area and the application of AI in South Africa is still in development in phase, as per the global
developments and like the rest of the world.
A few AI solutions based on machine learning and other technologies have been established and applied
within South Africa's public and private sector; however, the scope and usefulness of such technologies is
currently limited. There are principles that guide the need SA Action Plan as follows:
• AI should be part of the long-term wider National plan or strategy.
• Balancing act between enabling environment and required governance.
• The plan must be informed by a process to be undertaken.
• AI plan must be focused on Actions.
• AI is created by people for the development of humanity.
• AI can be used as a tool for communication and public awareness.
The following areas of focus will guide the adoption of the AI National Plan in SA:
Area of Focus Activities Timelines
Human Capital Development Basic Education (coding, robotics, Immediately

critical thinking etc)
Incentivizing AI skills initiatives Immediately

Skills knowledge and transfer (global Immediately
partnership)
Localized AI solutions AI Innovation Market Access 2025
AI venture investments and start-up 2025

supports
Supporting AI start-ups Immediately
Reinforce Collaboration with Innovation pipeline Immediately

Academia and Industry
Centralised Computing Power 2027
Distributed AI through Non- 2025

traditional data and Open Data
Regime
AI Hyper-scale Data and Large 2028

Language Model Centre
Capacitating Artificial 11 AI Hubs/Centers of Excellence 2025

Intelligence Institute of SA
(AIISA)
AI Data Markets and Revenue 2026

collection system
AI Regulation on AI Applications 2026
Data Computing 2025
AI industries 2027

AI copyrights/patents 2025
Protection Intellectual Property 2025
Unfair Competition 2025
AI Standards regulation 2025
Data Architecture Systems in EDUCATION: Integrating Data and Immediately

Key Sectors AI in education to align the
education system with labor market
needs and improve the student’s
journey
GOVERNMENT: Assimilating Data

and AI in Government to create a
smarter
and more efficient public sector
HEALTHCARE: Integrating Data and

AI into healthcare to increase access,
enhance preventative care and
accommodate growing demand.
ENERGY: Integrating Data and AI

into Energy in increase capacity,
enhance
efficiency, and develop adjacent

industries.
MOBILITY: Incorporating Data and

AI in Mobility to build a regional
hub, create smart city Mobility

technology, and enhance traffic
safety in cities
Create Data collaborative FDI and domestic investment, both 2025

Space in Africa from the government and the
private sector, would be essential to
support the development of the AI
support in the AfCFT
Global hub for AI skills creation

centre
AI Policy and Legislation AI Ethics and Governance 2025

Process
AI Maturity Assessment Tool which can be used to measure 2025

Framework for SA the level of maturity South Africa is
at, in term of AI
Set-up AI Ministerial Advisory Deliberate Targeted National 2024

Council or AI Panel of Experts Departments Network Industries
and Frontline services
CONSULATED SOURCES AND ACKNOWLEDGEMENT
Many publication sources on AI are widely available on the websites and published articles. Benchmarking
and best practice approaches would serve as a guidance towards the development of this draft document
which will be for discussion purposes solemnly.
South Africa belongs to several international bodies that outline the basic tenants of how AI can be
approached towards the fulfilment of national priorities such as the United Nations, International

Telecommunications Union, UNESCO, and OECD which would have been researched to ensure that proper
flow of South Africa’s approach is within the global prescripts.
It is important that the contents of the draft document are to guide the SA approach in relations to the
discourses taking place around the world in order to ensure that as a country we are not gravitating far
away from the global norm and standards.
Lastly, the SA national policy formulation process is based on extensive consultation as a basic tenant of
priority setting and this draft document is a measure to ensure that such outcome is always maintained.

National AI Government Summit Discussion Document

Uploaded by

Copyright:

Available Formats

National AI Government Summit Discussion Document

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

National AI Government Summit Discussion Document

Uploaded by

Copyright:

Available Formats

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING

Disclaimer – Content, Usage of the Draft Document:

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT

Alfred “Ali” Mashishi: Chief Director – ICT Forecasting and Modeling

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT i

Minister of Communications and Digital Technologies

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT ii

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT iii

Foreword by Minister of Communications and Digital Technologies ....................................................................... ii

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT iv

AI The International Landscape: AI Global Discourse, Trends and Drivers

-European Commission definition

A1.1 Introduction and Purpose

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 1

• Creation of policy and regulatory experiments

• Management of negative AI impacts on society and industry.

• Building an understanding of the AI technological possibilities.

A1: 2 Global Discourse: Politics of AI; Economics of AI; and Militarization of AI

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 2

• Labour market, and

AI Studies on its Impact on the African Continent

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 3

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 4

At a continental level, particularly in Africa, there is a strong emphasis on AI supporting sustainable

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 5

• Expanding connectivity to the Internet and Digital Infrastructure

• Enhance an ecosystem that promotes competitive and open markets

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 6

A1: 3 Investment in AI Technologies and Applications

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 7

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 8

- Data as a meaningful developmental tool.

- Digital skills and Job Impacts (Employment)

- Global Bodies Formation and Approaches on AI

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 9

- Funding and Investments – Protection of Intellectual Property and Patents

A1: 5 Main Types of AI Discourse

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 10

Diagram 1: Components of AI Ecosystem

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 11

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 12

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 13

A1. 7 Regional vs National AI Approaches and Frameworks

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 14

A1: 8 South Africa’s AI Conceptual Framework and Approach

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 15

• AI public skilling courses and raising AI awareness.

• Creation of AI sandboxes for testing AI localised applications.

• AI has the potential to deliver real value in South Africa.

Section B: AI – South African Landscape, Opportunities and Pitfalls

B1. AI Understanding within South Africa’s Context

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 16

B1. 1 Adoption of AI: SA AI Approach PC4IR

SOUTH AFRICA’S ARTIFICIAL INTELLIGENCE (AI) PLANNING: ADOPTION OF AI BY GOVERNMENT 17

• preparation through experimentation,

• private sector capabilities,

• technological clarity and commitment,

• government as an organizing player,