AI in Mining: Separating Reality from Fantasy

Read time: ~4min

ChatGPT has driven the latest wave of Artificial Intelligence (AI) hype globally, and mining is not immune to the hype cycle. Humans tend to react strongly to advances in new technology due to our innate aversion to uncertainty and unknown outcomes (Loewenstein, G. & Weber, E.U., 2013). The fear or excitement stems from the perceived impact that new technology will have on our daily lives, jobs, and overall well-being (Turkle, S., 2011). It's crucial for individuals to understand the potential benefits and drawbacks of new technology to reduce irrational fear or excessive excitement, both of which are counterproductive to moving our industry forward.

AI has the potential to greatly impact the mining industry when it comes to decision making. However, there are two extreme views of what AI can achieve in the industry, and the reality is likely to be somewhere in between. In this blog, we will explore the difference between these two views and what the reality of AI in decision making in the mining industry is likely to be in the coming years.

What we are currently doing without AI

Currently, decision making in the mining industry is primarily done by humans. Control systems control and react in accordance with specific programming. Data analysis is done in arrears and utilises many forms of mathematical and logical aids, but ultimately is still a human driven process. The majority of decisions are based on human experience, knowledge, and intuition, with or without underpinning data.

While this approach has served the industry well in the past, it has limitations. Humans are susceptible to bias and errors, and the sheer amount of data that needs to be processed can be overwhelming. It is generally accepted that the proliferation of IOT has vastly increased the available data, but we still only use a fraction of this data to make decisions. The ratio of data to cognitive capability is now many 1000s to 1, and still rising.

Finally, we have a well acknowledged “brain drain” in mining, of experienced people leaving the industry and fewer younger people coming in. As an industry, we rely on experience, and hire and promote on experience, and this experience is evaporating. The result is a kind of split personality that presents across organisations, with 1 side pointing to data in arrears, and 1 side having to make decisions at 3am based on a radio call from a graduate. Dropping the “AI” bomb into this atmosphere is often not conducive to effective decision making.

The two ends of the spectrum

At the extreme optimistic end of the spectrum, some people believe that AI will replace many human workers and make all operating, technical and financial decisions in the mining industry. This vision is based on the idea that AI is able to process vast amounts of data quickly and accurately, leading to better decision making. According to this view, AI will make all critical decisions in the industry, from resource allocation to equipment maintenance. At the pessimistic end, some people believe that no machine could ever replace their hard won and deeply technical experience and expertise. They believe that “technology” has promised so much over the decades of their working life and delivered so little, and the people building these “smart machines” have never been underground in their life, and therefore could not have any idea what is even needed to do their job.

What the reality is most likely to be in the coming years

In reality, AI is unlikely to replace any significant fraction of human workers in the mining industry in the next 10 years. While AI can certainly assist in the decision making process, there are many aspects of the industry that require human judgement and intuition, or a combination of physical activity and cognitive abilities. Additionally, there are still very real technical, cultural and practical barriers to the widespread adoption of AI in the industry, even if the underlying cognitive abilities were to suddenly leap forward.

What is very likely to happen is that AI will help mine workers by analysing large amounts of data from multiple systems and sensors, and providing near real time decision-making insights and recommendations. Many workers will start to receive advice from systems that are giving recommendations that conflict with their experience, and conflict with existing siloed expert systems. What the humans do with these recommendations, and whether the recommendations are right or wrong, or better or worse, and whether the AI and the humans can learn from this together and do better next time, is the battleground that mining will live in for at least the next decade. It may well be that humans will progressively move to learning and explaining, rather than deciding in the moment with incomplete information, and the shape of this human-machine symbiosis may alter the way we all live and work.

It is also very likely that mine owners will watch other industries race ahead with AI driven decision making, and they will say “me too”. Added to this, the drain of experienced people from the industry is very real and the demand for more mines to fuel the energy transition is weighing on everyone. While we cannot substitute an AI for a superintendent, neither can we substitute graduates in data science for that superintendent. What is needed is to help our mid level supervisors make better decisions, and to help fewer superintendents and senior engineers support more junior workers and contractors across more sites. This can only be achieved at sufficiently low risk with decision making that scales beyond the collective number of years working on site.

Conclusion

While the two extreme views of AI in decision making in the mining industry are unlikely to eventuate, it is clear that AI has the potential to greatly impact our industry, through changing the way humans make decisions. The adoption of AI in the mining industry is likely to be sporadic, and punctuated by failures, but as the technology becomes more advanced, available and widespread, it is likely to become an increasingly important tool for augmenting human decision making - as long as those humans want to learn and grow with the machines. With the help of AI, workers in the mining industry have the opportunity to make better and more informed decisions, leading to improved efficiency and productivity, and creating a more sustainable human strategy for the decades to come.

Reach out our team today if you would like to adopt a better AI & Digital Twin strategy today: