Power of AI: AI's big promises start to deliver for miners adopting new tech

The global mining sector is tapping into cutting-edge artificial intelligence to uncover faster and better ways to extract minerals from the ground and process them.

Advances in AI — primarily driven by machine learning that mimics human brains — are increasingly being used to monitor for costly equipment failures, find difficult-to-locate lodes, flag disruptions in the supply chain and tease out more valuable materials from the production process. While some in the industry say the most significant gains from AI are years from being realized, early adopters are capturing a clear competitive advantage, according to interviews and feedback from 11 mining industry executives and experts.

"AI is alive and well in mining — exploration, operations — everything that can help us take data and learn from it in a way that's quick and efficient and can lead to better decisions, quicker decisions," Mitchell Krebs, CEO and president of Coeur Mining Inc., a US-based gold-silver miner, told S&P Global Commodity Insights. "It's just a much more predictive world, and AI plays a part in that."

AI promise

As the industry faces a drought in discoveries of increasingly in-demand energy transition metals, AI-based exploration tools are gaining attention and funding. California-based KoBold Metals Co., for example, has the backing of billionaires Bill Gates, Jeff Bezos, Michael Bloomberg and Ray Dalio and is reportedly spending $150 million to accelerate its search for copper and cobalt at a project in Zambia using AI-based exploration tools.

The current process of discovering a new orebody, defining and moving it through processing could be more efficient, said Michael Ravella, chief innovation officer at Veracio, an AI-based mining technology company owned by mining and drilling services company Boart Longyear Group Ltd.. Digital technologies will allow for a granular, centimeter-scale understanding of a mineral resource that miners can leverage to optimize extraction and at points beyond in the supply chain, Ravella said.

"Digital sensing and AI can drive discovery. But once discovered, they can also move resources forward from discovery to resource to reserve to processing," Ravella said.

Processing boost

In combination with sensors, AI can predict how best to tweak the complex processing recipe at mines, according to industry experts and executives.

"I think the biggest opportunities are in processing. One reason is that, basically, there is a lot of data," said Tycho Möncks, a managing director and partner focused on the mining sector with the Boston Consulting Group. "And algorithms love data."

AI is already a virtual colleague in processing at many large mines, advising operators on how to liberate metals from complex and increasingly lower-grade ores. Executives said the programs do not replace workers but augment decisions at mines, especially in milling.

Seven of Freeport-McMoRan Inc.'s 10 concentrators have AI-enabled models that help direct input flows and other steps, said Cory Stevens, president of the copper-mining giant's mining service division. Concentration is one of the final steps in ore processing.

Freeport is rolling out similar AI-driven systems to heap-leaching operations, and, in a first for the company at an open pit mine, it is going fully autonomous at the Bagdad copper operation in the US. Amid a frenzy of mining activities, AI will help coordinate and decide how best to run the fleet, such as by judging when to charge batteries for increasingly electric-powered equipment.

Freeport expects to reduce sitewide human resources by about 10% at mines where it deploys automated haulage, Stevens said.

Likewise, miners have widely embedded AI in advanced grinding equipment, said Mikko Tepponen, chief digital officer and chief operations officer at FLSmidth & Co. A/S, a large mining equipment manufacturer and service provider. Tepponen estimated that machine learning has helped deliver 10% throughput improvements in semi-autogenous mills, which grind ore while using the same amount of power as non-AI-aided systems.

"If you think about some of the big producers of the world ... it leads to $100-million-dollar-plus yearly benefits," Tepponen said.

In AI-assisted concentrators, Freeport's metal recoveries can improve by a couple of percentage points while throughput improves by 10% to 15%, Stevens said.

"Over time, the operator starts to get the hang of it, and they're gaming it ... where they're trying to beat the AI before the recommendations are actually made," Stevens said. "And then the machine actually will learn from itself as well."

Permanent memory

AI is also proving useful as a keeper of institutional knowledge, industry executives said.

"It's a permanent record, and as time goes on, metallurgists might change, different plant managers might come through," Freeport's Stevens said. "But having that history is immensely valuable."

Along the same lines, Sydney-based Hevi AI is using AI to process large documents, such as equipment and safety procedures manuals. It makes them available to workers through a chat interface. Such advancements, the technology promises, allow miners to be more productive and efficient by giving them instant access to reams of information.

Elsewhere, AI is making quick inroads into preventive maintenance, where mining executives said capabilities will grow as sensor and data networks spread. It can detect subtle problems and double the life of some equipment, Freeport's Stevens said.

"That adds to availability and cost reductions and efficiencies that we didn't think were possible before," Stevens said.

One untapped area for AI to exploit is data sharing, executives said. Information related to safety and environmental metrics would be easy for industry to share more widely, as proprietary information may not be at risk, they said.

"We believe that's the next frontier," said Perry Zalevsky, senior director of industry at Aveva Group PLC, a UK digital infrastructure provider to the mining sector. "We think the next thing is this community, this sharing."

AI resistance

However, resistance to adoption, a lack of data and the unknown risks around nascent technology present challenges for miners looking to deploy AI or automation, some executives said.

"There will be a time and a place," said Peter Marrone, the Yamana Gold Inc. founder who sold the company in 2022 and is now chairman and CEO of Allied Gold Corp. "AI could be very helpful. But we're still a long way off from that becoming the ordinary course."

AI and automation can reduce costs for miners, but some companies must also consider the employment they offer as a social license to operate.

"In the West, people are using technology to displace labor," Mark Learmonth, CEO of gold miner Caledonia Mining Corp. PLC, said in an interview. "In a place like Zimbabwe ... one of the attractions for the government is to create jobs. If we were to suddenly turn around and say, 'We are going to automate the entire mine and all those jobs,' that wouldn't go very well."

Meanwhile, executives said AI will mean hiring for a new and highly in-demand skill set, and they must compete with Silicon Valley behemoths.

"Thanks to the digital revolution, we now find ourselves in a global war for talent with every other industry," Dean Gehring, gold-giant Newmont Corp.'s executive vice president, chief integration officer and acting chief technology officer, said in an email.

And mining companies and their employees will have to overcome their own conservative natures.

"The situation they are facing is, first of all, with themselves: that basically there is someone that knows better, and that someone is a computer," Boston Consulting Group's Möncks said.

Looking further into the future, some executives expect AI will allow miners to integrate self-operating equipment, like robots, deeply into their processes. At first, AI-powered robots may take over more dangerous tasks, such as checking out rock falls and gas leaks, Stevens said.

"I think it's going to be a gradual sort of adoption," Stevens said. "But then it's going to be an exponential curve."

S&P Global Commodity Insights produces content for distribution on S&P Capital IQ Pro.