In February 2024, KoBold Metals, a Silicon Valley-backed exploration company, sent shockwaves through the mining world with its announcement of a potentially massive copper deposit in Zambia’s Mingomba region, writes Dr Nicolaas C Steenkamp.

Wirestock on Freepik

Wirestock on Freepik

The Mingomba discovery, touted as the largest in the country’s history, raises not only economic hope but also questions about the role of Artificial Intelligence (AI) in modern mineral exploration and its implications for sustainability and social responsibility.

 

Historic data and AI fusion

KoBold’s approach is unique in its reliance on AI and data analytics. Instead of traditional exploration methods heavily reliant on on-the-ground surveys and drilling, they utilise a “digital first” approach. This approach involved utilising an amassed vast repository of historical data, including geological maps, geophysical surveys, satellite imagery and even historical mining records from decades past. This rich tapestry paints a comprehensive picture of the subsurface potential. This raw data, often inconsistent and incomplete, undergoes rigorous cleaning and standardisation to ensure accuracy and compatibility with AI models.

Machine Learning Algorithms were developed by KoBold, including convolutional neural networks and probabilistic modelling, to analyse the processed data. These algorithms extract patterns and predict the location and extent of potential mineral deposits. Despite relying on AI, human expertise remains crucial. Geologists familiar with the Central African Copperbelt interpret the AI outputs, incorporating contextual knowledge and understanding of mineral formation processes.

In the case of the Mingomba discovery, this data-driven approach proved highly successful. By analysing historical data across a wide area, KoBold identified geological anomalies suggestive of large-scale copper mineralisation. Targetted drilling confirmed their prediction, revealing a high-grade copper deposit estimated to be worth billions of dollars.

 

Hunting for elephants or revolutionising exploration?

Sceptics have however doubted the uniqueness of the discovery and credit “hunting for elephants in elephant country” as a key driver of the discovery. This is due to the fact that the deposit is located in proximity of known mineralised areas and that the area has been drilled and extensively described by exploration geologists.

The approach used by KoBold prioritised precision over randomness. Unlike blind prospecting, KoBold’s approach is data-driven and precise. By analysing vast amounts of information, they target specific areas with high probability of mineralisation, significantly reducing the exploration footprint and increasing efficiency.

Traditional exploration often involves extensive land disturbance and resource depletion. KoBold’s method minimises these impacts by relying on existing data and targeted drilling. The social license to operate is also considered, where community engagement is required. KoBold emphasises social responsibility and transparency, engaging stakeholders throughout the exploration process.

While historical methods might resemble “hunting for elephants,” KoBold’s approach represents a shift towards a more data-driven, sustainable and responsible exploration paradigm.

Exploration corebox.

Exploration corebox. Supplied by Dr Nicolaas C Steenkamp

 

 

Challenges and the road ahead

Despite its promise, KoBold’s approach faces challenges. The approach is highly data driven, which requires access to comprehensive and reliable data, which remains a barrier in many regions. It is also a very new exploration method that has largely only been tested theoretically, thus algorithmic bias could still occur. Ensuring that AI models are free from biases which could lead to skewed results, is crucial.

Balancing economic benefits with minimising social and environmental impact requires careful planning and community engagement. The company remains optimistic about overcoming these challenges and transforming the mining industry. They are currently seeking partnerships and exploring collaborations with academic institutions to further refine their AI models and address sustainability concerns.

 

A new era for mineral exploration?

KoBold’s Zambian copper discovery showcases the potential of AI-driven exploration to reshape the mining industry. However, ethical considerations and responsible practices are equally critical. As this technology evolves, ensuring transparency, collaboration and environmental stewardship will be key to unlocking the true potential of this “digital hunt” for valuable resources. While it is too early to declare a revolution, KoBold’s success stands as a testament to the potential of AI in transforming traditional exploration methods and ushering in a more sustainable and responsible future for the mining industry.

The need for skilled geologists to interpret results also remains crucial, which requires years of hands-on experience in the field and not purely relying on computer generated models. The ultimate proof of the success however remains to be determined as ground truthing is done and the exploration programme validates the claim.

Dr Nicolaas C Steenkamp is an independent consultant, specialising in geological, geotechnical and geometallurgical projects and mining project management. He has over two decades of industry experience with global exposure. (ncs.contract@gmail.com)