As we transition from our May editorial theme of smart mining to June’s which centres on people, I’d like to reflect on a topic that bridges these two worlds – generative AI. Like it or loathe it, this technology is here to stay, both in our personal and professional lives. Why? Because it provides a way for individuals and organisations to get closer to data.
As Deloitte explains in its 2024 Tracking the Trends report, foundation models are what differentiate gen AI from traditional AI. These are complex learning models that are trained on a broad dataset and can be adapted to solve various problems.
The team states: “Many foundation models today are in the text domain and are driven by vast quantities of available training data. However, gen AI can create artifacts across various modes, including code, images, video, audio, and 3D models… Increasingly, the possibility for different modes and even multimodal models could both disrupt and drive step changes in productivity across a range of functions.”
In the near term, even today, these capabilities are being harnessed in back-office functions as virtual assistants or ‘copilots’ for mining software developers and data scientists, as McKinsey explains in this article. But it’s the long-term impacts which could prove the most interesting.
According to McKinsey, using custom-built models, mines could eventually create “libraries of maintenance manuals, historical work orders, procedures, tooling inventories, and parts databases” that technicians can search and apply verbally and almost instantly while in the field, helping to streamline work and increase reliability.
Further into the future, Deloitte believes that miners could use gen AI’s ability to simulate, model, and generate data-driven insights to support supply chain resilience and optimisation, and to rethink processes like mine design and delivery. The possibilities seem almost endless… But, as we stand on the cusp of this potential explosion in capabilities, it’s pertinent to ask ourselves: at what cost?
The computing, infrastructure and hardware requirements for developing and training gen AI models are staggering – companies are quite right to be wary of the capital costs involved. And that’s before we consider powering them. Did you know that creating a single image using gen AI can consume as much energy as is needed for a full smartphone charge? And what about the ethics surrounding data collection, processing and usage?
If mining companies are to harness gen AI to its full potential without undermining their sustainability commitments or corporate integrity, then now is the time to think about these things.
What are your thoughts on gen AI and its proliferation in mining? Drop me a line and let’s chat.
Carly
Start a discussion
Piqued your interest? Share this with your colleagues, friends, family…