UJ’s Prof Tshilidzi Marwala stresses the need to integrate AI in SA’s mining industry


UJ’s Prof Tshilidzi Marwala stresses the need to integrate AI in SA’s mining industry


Publishing Date: 9/3/2018 9:00 AM

The Vice-Chancellor and Principal of the University of Johannesburg (UJ) and the co-author of the book Condition Monitoring Using Computational Intelligence, Prof Tshilidzi Marwala recently penned an opinion piece, Mining in the Fourth Industrial Revolution, published by the Sunday Independent, 09 September 2018.

Mining in the Fourth Industrial Revolution

A few weeks ago, we woke up to the news that 21 people died in the Sibanye Gold Mines. This is approximately 50% of all fatalities in South African mines this year. According to the USA Department of Labour statistics, the annual number of fatalities in coalmines in that country has dropped from 3342 in 1908 to 15 in 2017. In South Africa, mining fatalities have dropped from 290 in 2000 to 82 in 2017. The sad reality is that despite these improvements, our mines have not reached the safety levels that our country deserves.

Growing up in Venda in Limpopo Province, one of the ways in which parents motivated their children to study was to scare them by threatening them that if they do not study, they will end up working mugodini. Mugodoni literally means working inside a hole. By this they meant working in the mine. Given these fatalities, it seems this tactic of threatening people to work in the mines was effective because very few people from Venda ended up working in the mines.

The truth is that despite the impressive gains that have been made as a result of industrialisation, the mining industry is still stuck in the first and the second industrial revolutions. The first industrial revolution gave us mechanised means of production. The second industrial revolution gave us the assembly line and electricity, while the third gave us electronics and is driven by the semiconductor industry. Now the fourth industrial revolution is automating production at such a rapid pace that human beings are increasingly being rendered irrelevant in the workforce. The fourth industrial revolution is driven by advances in artificial intelligence, block-chain technology and biotechnology. If we are to measure the industry's citizenship in the fourth industrial revolution by measuring how many semiconductor devices – i.e. phones, computers, and intelligent robots – are used for production, and then the mining industry is lagging behind other industries such the automotive, defense and manufacturing industries.

To understand the slow pace of the advancement to the fourth industrial revolution by the mining industry, we need to understand its political economy. In this regard, mining companies acquire mining rights from our Department of Mineral Resources. Mining license is a right to mine a particular mineral at a particular location for a particular period. To get this license the mining company must do environmental impact assessment as well as consult with the local community. These companies then put the mining infrastructure, employ people and operate the mine. The pact that government reaches with these companies is that mining companies employ people, pay taxes and pay royalties to government.

In game theory, which is a mathematical field that was popularised in the movie The Beautiful Mind, there is a concept called a Zero Sum Game. This concept means that the gain by one party is always a loss to the other party. The mining industry is a zero sum game because the gain of minerals by these companies is necessarily a loss to the minerals balance sheet of a country. What government hopes is that employment, taxes and royalties compensate the loss of mineral resources by a country. The other reality is that the employment relations in the mining industry have not changed much in the last 100 years. Similar to hundred years ago the wages in the mining industry remain low and the work environment is not yet 100% safe.

Now why has this industry not invested into the technologies of the fourth industrial revolution? One reason is that the mining environment, especially in deep mines, is hostile to these technologies. The temperatures, humidity and noise levels are very high. Therefore, a semiconductor device such as an artificial intelligence powered robot does not last. One thing that is crucial in semiconductor-based technologies is temperature control. Our computers have to be cooled down continuously by internal fans otherwise, they will melt. Data centres consume large amounts of energy because they have to be cooled down. When we put such semiconductor devices, in mines where temperatures can rise up to 55oC, the amounts of energy needed to cool them down is high and the relevant technology is not yet available.

The second reason why the mining industry has not moved fast enough towards the fourth industrial revolution is humidity. In Metallurgical Engineering, we learn that humidity causes materials to rust quickly. We also learn that humidity makes devices unpredictable because humidity levels influence the performance of machines. The third reason is that the environment underground is messy and therefore a robot that has been designed to operate on a factory floor cannot work very well underground. This is because there has been considerably more resources invested in factory robots than underground robots. The fourth reason is that sensors that are needed to power these robots are designed for over the surface environment rather than the underground environment.

Given all these limitations, how do we reconcile mining and the fourth industrial revolution? Given the pact that government gives companies the right to mine and companies in return employ people as well as pay taxes and royalties, how do we handle the fact that the fourth industrial revolution will decimate employment? How do we reconfigure taxation, given the fact that the mining industry is an extractive industry? How do we deploy this technology to ensure that we reach zero fatality rates as matter of course within the next five years? 

To attempt to answer these questions, we need to understand how the fourth industrial revolution is generally implemented in the mining industry. The ways in which the fourth industrial revolution is implemented is through processes, systems and infrastructure. In this regard, this industry needs to deploy processes and systems to monitor the movement of people, machines, rocks and slopes underground. The technology of the first industrial revolution gave us tools to analyse the movement of rocks through the work of the Russian Engineer Stephen Timoshenko. We can integrate this technology with artificial intelligence so that we are able to predict movements of rocks and slopes and protect people as well as increase production. The cameras that are deployed underground should be augmented with artificial intelligence so that they can predict movements of people, slopes and vehicles to protect lives and increase productivity. The energy consumption should be reduced by deploying energy harvesting technology underground. The vehicles underground must be autonomous, intelligent and connected to other vehicles, through the internet-of-things technology, to minimize accidents and maximize production. We need to invest in understanding the psychology of people working underground and their interactions with machines and technology within the context of a hostile environment.

Going back to people, we need to understand the overall impact of automation in the mines and its interaction with technology. The first industrial revolution created trade unions and united the working class to fight against exploitation. Yuval Harari, in his book 21 Lessons for the 21st Century, notes that the fourth industrial revolution is making the working class irrelevant in production and that it is easier to organise against worker exploitation (in the first industrial revolution) than worker irrelevance (in the fourth industrial revolution). It is also expected that the reduction of people in the mines will result in less tax collection from this sector. This will increase social burden which will be borne by society and government. A special tax regime specifically for extractive industries will have to be developed. Currently mining companies pay corporate taxes and royalties to government and the important question is whether these are enough. We need to invest in our educational institutions to develop technologies suitable for the underground as well as understand the psychology of people working with automated machines of the fourth industrial age.

The views expressed in this article are that of the author/s and do not necessarily reflect that of the University of Johannesburg.