Mining’s AI transformation needs to be done right or it’s just a distraction, McKinsey expert warns

Artificial intelligence will be crucial in breaking the mining sector’s decade-long productivity standstill, a McKinsey expert has predicted, but done wrong it’ll be little more than a distraction.

Perth-based partner Cris Cunha has been consulting the resources sector about AI for the past decade or so, and is studying for a PhD on where the two worlds intersect at the University of Western Australia.

From staying across the latest applications, to ingraining AI throughout an entire business, he’s learnt firsthand how much of a help the technology has become, but also how much of a hindrance.

Mr Cunha said this was particularly the case with the advent of generative AI — a new iteration popularised at light speed over the past 12 months by ChatGPT — that can be used to generate audio, videos, pictures and written work on cue.

Autonomous haulage and remote operations have been sizeable leaps for industry, but he believes embracing AI will be critical if mining companies are to shift the dial on languishing productivity levels that he says have been “flat since 2015”.

“Mining cannot take the next leap of productivity that it needs without AI, period,” he said.

“Generative AI, which is huge advancement in AI over the last year . . . has reinvigorated the focus of AI in the mining industry. However, my biggest word of caution is that if it’s not done right . . . it’s not going to to move the needle as we would expect.”

He said there was a risk that rolling out generative AI without a clear focus could cause “distraction” and companies might forget to use AI for business fundamentals, which was ultimately “all about extracting value from a mineral in the ground”.

Mr Cunha said he was seeing more unsuccessful AI experiments in the industry than successful ones and only about 20 per cent were flowing through to a company’s bottom line.

And it’s not as simple as rolling out a algorithm loaded chatbot, either.

“If you think about using AI to predict failures for a certain mobile equipment, let’s say a haul truck, if you do not fundamentally transform the way you do maintenance . . . all that you’re going to have is a prediction that something is going to fail and you can’t do anything about it,” he said.

“In order to scale and sustain AI in the mining organisation, you cannot think about doing AI projects, but it’s all about rewiring the whole way that the mining company needs to work going forward.”

Mr Cunha is a partner at QuantumBlack, McKinsey’s forward-looking AI consulting arm.

One of the firm’s big successes has been rolling out AI systems for US copper major Freeport-McMoRan, which was reluctant to navigate lengthy approvals processes for a new mine or fork out for brand new infrastructure to boost production.

“Freeport McMoRan . . . developed an artificial intelligence system to regulate the operating parameters of the mineral processing facility, they managed to deliver $US300 to $US400 million out of that system in one year. And it saved them the need to construct a whole new concentrator,” Mr Cunha said.