Programme Director, Ms Leanne Manas;


Ms Michelé Schliesser, Chief Information Officer, South African Reserve Bank;


Mr Seaparo Phala, Chief Information Officer, Department of Arts and Culture;


Various speakers for today;


Distinguished guests;


Ladies and gentlemen:




I would like to thank the organisers of this event for inviting me to be a participant in this seminar.  The technological changes that are taking place in our country and around the world require all of us to pause and reflect on the impact these changes are going to have on every facet of our lives, and, more importantly, on the different areas of work.




Those of you who are familiar with social media would have heard of or used a social media platform called WhatsApp.  Through this platform we are able to communicate with each other through direct messages and voice calls.  This platform is owned by Facebook, which is domiciled in the United States.




Tencent, partly owned by our very own Naspers and domiciled in China, has created a similar platform called WeChat.  It is the most popular social networking platform in China.  I have a Chinese friend with whom I communicate using WeChat.  I have realised only recently that when he communicates with me through this platform, he writes his messages in Mandarin, and the platform translates his messages into English; and similarly, that when I write my messages in English they are translated into Mandarin.  The platform uses an artificial intelligence engine to translate between the two languages.  The AI engine has the ability to learn from data over time, and as a result this engine has become very good at translating between these two languages.




Locally, Google also has an artificial intelligence platform that can translate from isiZulu to English and vice-versa.  When it translates from English to isiZulu, the machine does very well.  However, translation from isiZulu to English is very poor, and the same is the case with other African languages such as Swahili.




The reason for this, as Professor Tshilidzi Marwala noted in a recent article, is that while the people who built the machine are proficient English speakers, there are no isiZulu speakers participating in creating these machines. These machines learn from data; if they are not provided with data, they will not be able to learn.  What makes the WeChat platform better at translating is that Mandarin speakers developed the platform, including the AI machine. This means that as South Africans we should collect our own data or information, manage it ourselves, and develop our own AI systems.




Why are these platforms important?  More importantly, why are these platforms important for the topic we are discussing today?




The AI engines used in these platforms can be used in various areas for record keeping and information management.  In addition to facilitating communications, the other thing these platforms do – and which many of us are not mindful of – is collect data and keep records.  They collect data about our whereabouts, what we say to each other, our market preferences, our pictures, and many other things that we share through these platforms.  Actually these platforms can tell you a great deal about a person, a community and a country, and can influence decisions taken regarding them.




South Africa has 11 official languages, and this means that the records we keep can be written in any of these languages.  To increase the accessibility of these records for all South Africans, we get translators to translate them into other languages.  A good example of this is provided by our courts.  English is the only language of record in South African courts, but this doesn't mean that people can't use other official languages.  AI machines could be used for real-time translation and for more reliable record keeping.




In order to improve the quality of health care in South Africa, the government has opted to implement the National Health Insurance (NHI) policy.  The scheme is aimed at ensuring that all South Africans get access to quality health.  For the NHI to work efficiently, it must be able to register and track patients.  The Department of Health, the Department of Science and Technology and the Council for Scientific and Industrial Research (CSIR) have developed a patient registration system that could be used for this purpose together with electronic health records.




Electronic health records allow for the capturing, storing and sharing of a patient's personal as well as medical information.  This enables health care providers to get access to patient data at points of care in different geographic locations, and to make more informed decisions about health care provision on the basis of this data.




Within our health care system in South Africa there are traditional healers.  These are practitioners of traditional African medicine, and they are regarded as indigenous knowledge systems practitioners who treat patients.  Because most traditional healers use their respective vernacular languages, an artificial intelligence language translation system would be useful for ensuring that the information collected by traditional healers from their patients was accessible at any point of care, to assist other professionals in providing quality health care.




Artificial intelligence can also perform data analytics for predictive analysis and pattern recognition.  This can be of great assistance in the early detection of epidemics and other outbreaks of infectious diseases.




Data provides significant insights into government operations when it is timely, correctly collected and analysed.  The maturity of artificial intelligence techniques in providing insights from data has created a new phenomenon, in which data is seen as the "new oil".




Most government departments and agencies have drawn up strategies for harnessing the power of data.  Central to this is the ability to implement effective determination, protection and authentication of individuals.


The Department of Home Affairs (DHA), which is charged with securing the identities of citizens and maintaining a national population register, has through its modernisation programme developed the smart ID card in partnership with the CSIR.  The smart ID card allows citizens to participate in the digital economy with complete trust of the identity mechanism during a transaction.




The smart ID card stores a very small segment of the National Population Register and the Home Affairs National Identity System (HANIS) in a very secure manner on the chip, only allowing trusted systems to communicate with the card.  This makes the smart ID card an invaluable asset in terms of identity data management.  Biometric features stored on the smart ID card provide unparalleled citizen verification capabilities which do not require real-time connection to the DHA's back-end systems.  The smart ID card was designed to form the basis for citizen identity in digital government, with the capability of hosting several government cards, such as electronic driver's licences and electronic health record cards.  The smart ID card allows other government departments or agencies to electronically read a citizen's record when they apply for a service – such as a grant from the South African Social Security Agency – thus removing the need for manual data capture which often leads to typographical errors.




What I have cited are just a few examples of how AI can be used in record keeping and information management.  It is evident that as we introduce more smart data collection systems, we will also need more advanced data storage and management systems. The speed with which we now collect data and the amount of data we are collecting have resulted in what is called big data.  It is not enough to have the data; to derive any value from the data, it is critical to possess big data analytics tools.  Herein lies the importance of artificial intelligence.




AI can be used for record keeping and information management in the following ways, among others:

  • Machine learning algorithms, which are a subset of artificial intelligence, can comb through unstructured text and data and learn about the format and the content as they go


  • AI systems can assist in the classification of records and information once they are taught the various classes of data.


  • Machine learning can be used to eliminate documents that are no longer needed or that are duplicates.


  • Another potential application of AI is in compliance; an AI machine can assemble all the relevant documents needed for an audit within a matter of seconds.


  • AI can assist with data quality by helping to overcome common human mistakes when entering data, such as formatting errors, misspellings, misplaced information and other glitches.




AI machines will lead to improved record keeping and information management in the form of increasing data availability in real time, interoperability of data platforms across different systems, and appropriate packaging of data and data analytics for end users that will be value transformative to every sector's performance. It is important that the AI systems we develop to assist with record keeping and information management are such that they can –


·           be as transparent as possible, for fairness;


·           be auditable, so that the decisions taken with their assistance can be reconciled with systems of accountability;


·           be incorruptible, which means they must be secure, so that all of us can have confidence in their usage;


·           be predictable, in that the same inputs must produce the same output.




Artificial intelligence systems are going to change our lives in many ways.  While some of the changes will be unpredictable, it is important for us to study the ways in which things are going to change and to prepare accordingly.  The Department of Science and Technology, through its entities, is already doing a lot of work in this regard, and this will be supported by our new White Paper on Science, Technology and Innovation.




There is currently a lot of conversation in our society about the various elements of the Fourth Industrial Revolution (4IR) and its likely impact on the country.  Labour is concerned about job losses, individuals are worried about privacy, Members of Parliament are concerned about how to legislate and regulate.  Recognising the huge amount of work that needs to be done, the questions that need to be answered, and most importantly, the opportunities that need to be seized, the President announced the establishment of the Presidential Commission on the Fourth Industrial Revolution.  The commission will help the country respond comprehensively to the changing environment, including how AI can help us build a better country, and I hope that many of you will collaborate with the commission.




As I conclude, I wish to indicate that we will be launching a 4IR centre in partnership with the World Economic Forum (WEF) at the CSIR.  Through this centre we will be looking at building partnerships with various countries as well as the private sector, as part of ensuring collaboration and a better response on what needs to be done.



I thank you.