While many might have obscure ideas about what artificial intelligence (AI) is, owing to certain dystopian books and movies, the AI revolution together with the Fourth Industrial Revolution, is a long-term process. It will take between 20 and 50 years to reach the heights its most ardent proponents yearn for.

This will give humanity, for the first time, a social system where the underclass does not exist, explains Eastern Cape-based Coega Development Corporation. Instead, computers and intelligent devices play the role of the underclass.

The AI revolution, therefore, is a megatrend that will take decades to unfold, and it has to be seen and managed as such. There are two key aspects, namely machine learning (including deep learning) and blockchain technologies.

The “machine” in machine learning is not a robot with servos and steel, but mere software. Software that may one day become ‘colleagues’ or take over tedious tasks. Machine learning is becoming an accessible “commodity” that is attained using computing algorithms, using data science.

A key question is what can government do to manage this megatrend. The first point to bear in mind is that the business of government in South Africa is mostly governed by the Public Finance Management Act. The processes in government are thus ‘fixed’ and tend to result in a form of inaction, which may be malign and lead to misuse, abuse and malfeasance within government as the custodian of public funds.

Machine learning can be brought to bear in terms of robotic process automation (RPA) and specifically the RPA for government and public sector referred to as RegTech. This RPA will be combined with another AI technology, which is, natural language processing. This will be in the form of chatbots that will be used mainly for service delivery. This could release billions of rands that are at the moment either misdirected or misappropriated, to drive economic growth and social cohesion.

Another way the South African government can leverage AI is through policy. Many governments globally such as Germany and China as well as many states in the US have innovative policies to promote the industry of AI, which itself is already worth billions of US dollars, and to ensure that AI propels industrialisation, with special economic zones (SEZ) being one of the drivers in this process.

South Africa would do well to copy the above-mentioned nations’ policies. It could also pivot its reindustrialisation strategy using AI as an instrument.

Industrial innovation remains an urgent matter. SEZs should not only foster new driving forces, but also transform and upgrade old driving forces to achieve coordinated development of both old and new driving forces.

The Internet of Things will be able to greatly reduce intermediation, monopolisation and information asymmetry, as well as company costs. It will increase the need for the development of smarter cities around SEZs.

The corresponding intelligent services will be provided in the process of making public services – including urban medical treatment, health and environmental protection – more intelligent.

It will also generate more need for intelligent production. The traditional enterprises in the SEZs will be upgraded and transformed into factories manufacturing or using AI.

SEZs will need to adjust and be more closely integrated into the global division of labour in the industrial value chain so as to build more extensive and deeper economic trade contacts between international and internal value chains.

AI has the potential to release humans to perform more “engaging tasks” and to deal with repeated and mundane tasks.

However, there are ethical considerations attached to AI. With RegTech, as an example of AI that is deployed in public sector procurement and is able to adjudicate and select successful tenderers, the processes, that is the algorithm that is used to select, must be transparent and understood by the data scientist who develops it, management in supply chain, the internal and external audit functions as well as stakeholders.

It is also important that the associated so-called deep learning neural networks are not deployed for malign purposes or to do harm. These challenges also give an indication to the new skills that need to be deployed in the area of AI, inter alia in data science, management, and monitoring functions such as analytics and audit functions.

It may be that the benefit of the digital transformation accrues to those who possess and/or use AI in an intensive manner. They will be joined by those who are skilled in the new digital technologies and those individuals who possess flexibility to embrace the productivity impacts of digital transformation in their field of expertise.

The rest of the workforce will be left behind in the sense that, even though there will be reduced unemployment, they will perform jobs that are of a lower economic value that are in no feasible manner, either due to their nature or a cost-benefit calculation, suitable for AI.

However, this should not be a major concern as the opportunity for up- and re-skilling in an effort to adapt to AI remains one of the important cornerstones of Industry 4.0.

South Africa is a developing country and with developing countries desirous of industrialisation to attain developed status, the choice for the country is either to use the productivity-inducing AI technologies or seek to follow the labour-intensive economic sectors such as textile and garment making.

These sectors, which may be the last to be affected by digital transformation, have the downside of being economically low in value owing to low productivity.