The mine of 2030 combines aspects of the interconnected mine, AI and the intervention of humans to ensure accurate insights that can be communicated in a manner that makes sense to everybody involved in the running of the mine. Johan Oelofse, engineer at multi-disciplinary engineering company DRA Global shared this view with the audience through his presentation at the 2025 edition of the 4IRinMining Seminar on 24–25 July at the Wits University Education Campus in Johannesburg.
“Interconnected systems help us unlock the mine of the future. However the use of AI in the mining environment is not about replacing people with technology, because the experience they have collected over the years working on mines is useful. Ultimately, the aim is to get to the right decisions quicker and turn all the generated information into useful insights,” he points out.
ALSO READ:
Reimagining mines through tech adoption
Building an interconnected mine involves a journey that includes adopting a digital mindset, digital connection, digitalisation and digital transformation. This is a complete journey from start to finish, but to get the maximum value from the process you need to know and understand the problems you are solving for, says Oelofse.
“The interconnected mine looks through the entire value chain and puts all the information on the dashboard for the operator. Onsite, you need to make sure that there is proper production, efficiency, safety and also look after the environment and the community, and the interconnected mine gets this juggling act right.”
However Oelofse says when an interconnected system is applied, we must realise that one mine is different from the next. “We need to provide context to enable technology to help us. This will enable us to take a specific approach to solve the problem. People are very important in the whole equation and need to be trained to know how to utilise the technology and how to use the results, because people, processes and systems need to work together. We need to know what technology, what system and what information needs to be pooled from where and which data sources need to talk to each other so that the problem can be contextualised and we get the right information. Then we can design the correct architecture.”
AI in mining
AI is useful as it can do the hard work, of pulling information from many different sources and combining it into one, for you. It can even write reports, as well as codes, for you. However, to achieve the mine of 2030 we need to train AI to work specifically for a given mining environment so that it can give us contextualised information, says Oelofse.
“Mining operations produce big databases. Intelligence comes from varied sources including IoT (Internet of Things) devices, alarm systems, information from decisions people have made. All this information needs structure and contextualisation to enable you to get the correct information and results.”
While generative AI is able to generate immediate responses to queries, AI does not necessarily work at the same speed in the mining environment, says Oelofse.
“For AI to be impactful in this environment, we need to take time to properly integrate it with the existing systems to have the opportunity to learn the behaviour of the overall mining system over time. Eventually it will start giving us information good enough to use. This happens after we allow it to learn through the process and experience and repetition, which takes time. This is where human intervention comes in to humanise the intelligence and then feed it back into the system to teach it. Unfortunately this is one of the pitfalls, having to wait for some time for it to start giving you useful insights, after going through the effort of constantly feeding it the right information.”
“Moving from where we are today to 2030, while data is king, we need to enable the operators to contextualise it so that we can draw specific insights from it because they are now able to see things they were not able to see before. The same decisions they make can be converted into information to feed back into AI to teach it to improve the information it generates.”
