Foundations of AI in Mining White Paper V2
Published: June 13,2024This project updates the Foundations of AI white paper to reflect advancements in machine learning (ML), edge computing, and federated learning, a decentralized approach to training machine models. The paper provides a structured approach to AI adoption in mining, emphasizing iterative implementation, quick wins, and learning from failures. It clarifies AI methodologies such as supervised and unsupervised learning, deep learning, and predictive modeling while addressing concerns like bias, data quality, and scalability. Additional papers will showcase real-world AI applications, including automated geological modeling, predictive maintenance, and optimization of ore recovery.
BACKGROUND
The mining industry is increasingly using artificial intelligence (AI)-based innovation as a tool to optimize processes, enhance decision-making, derive value from data, and improve safety. That said, levels of maturity vary throughout the industry resulting in the need to learn more about what AI is and how it can be applied to a specific mining operation. As a result, mining operations still face many challenges with implementing AI, such as establishing a data infrastructure. Many mining stakeholders also have concerns about how AI will affect the workforce along with uncertainty of the risks to committing to a multi-year project that can potentially fail. Additionally, rapid advancements in AI, such as generative AI, multi-agent systems, and causal AI, demand an updated framework to ensure relevance and applicability in today’s mining operations.
This paper will cover the basics needed to cut through the hype, address concerns, and clarify what methods are useful. It provides a realistic approach to moving forward with AI projects; it also acknowledges the importance of implementing an iterative process that focuses on quick wins and learning from failures and challenges in order to build the foundation needed for long-term success. This tool aims to help mining companies build the necessary foundation for planning for and implementing AI solutions within their businesses.
PURPOSE
The purpose of this paper will be to provide an understanding of basics needed to address concerns and clarify what methods are useful when implementing AI in mining. It will provide a realistic approach to moving forward with AI projects and acknowledges the importance of implementing an iterative process that focuses on ‘quick wins’ and learning from failures and challenges in order to build the foundation needed for long-term success. This paper will aim to help mining companies build the necessary foundation for planning for and implementing AI solutions within their businesses. The revised paper will include comprehensive guidance on both foundational and cutting-edge AI technologies relevant to mining operations, including generative AI, agentic AI with multi-agent systems, and causal AI, to meet current and future industry needs.
The objective of this project is to revise the previous version of the Foundations of AI white paper to account for technological advances such as machine learning models, edge computing, federated learning, generative AI, multi-agent generative systems, and causal AI. The white paper will focus on providing an understanding of basic information to address concerns and clarify what methods are useful when implementing AI in mining.
In addition to the white paper revision, there will be an ongoing initiative to develop research findings and papers providing specific examples where AI has been implemented in mining.
IN SCOPE
The scope will focus on revising the existing white paper to account for machine learning, edge computing, and federated learning, and newly added topics: generative AI, agentic AI with multi-agent generative systems, and causal AI. (The revision will also briefly acknowledge symbolic AI’s historical role and potential in hybrid approaches like neuro-symbolic AI.) In addition, potential papers following the revision can include the following:
- Research Finding: Edge AI in Mining
- Research Finding: Federated Learning for Predictive Maintenance
- Example 1: Autonomous Drilling System
- Example 2: AI-Powered Exploration
- Quantum Computing in Mining
- AI-Driven Exploration Techniques
- Machine Learning
- Deep Learning
- Reinforcement Learning
- Natural Language Processing (NLP)
- Computer Vision
- Federated Learning
- Large Language Models (LLMs)
- Large Action Models (LAMs)
- Neural Networks (NNs)
- Graph Neural Networks (GNNs)
OUT OF SCOPE
This white paper and the research findings and papers are not an implementation of AI guidelines and will not refer to specific company names and their technologies.
TABLE OF CONTENTS
1 Introduction
2 Background
2.1 Getting Started in AI
2.2 The Big Picture
2.3 Emerging Trends of AI in Mining
2.4 The Role of People in AI
2.5 Success Factors
3 The Transition to the AI-Enabled Mine
3.1 AI in Mining Maturity Model
3.2 Level 1: Basic
3.3 Level 2: Foundation
3.4 Level 3: Integrated
3.5 Level 4: Decision Supported
3.6 Level 5: Automated
4 Implementing AI
4.1 Choosing an AI Project
4.2 Structuring an AI Project
4.3 Building an AI Team
4.4 Benefits of an AI Project
4.5 Challenges and Risks of an AI Project
5 Transformative Application of AI in the Mining Industry
5.1 Enhancing Mine Site Efficiency and Productivity
5.2 Safety and Geotechnical Risk Management
5.3 Geological Modeling and Exploration
5.4 Advanced Mine Planning and Scheduling
5.5 Sustainability and Climate Resilience
5.6 Workforce Empowerment and Knowledge Management
5.7 Robust Data Infrastructure and Interoperability
WHAT'S HAPPENING
- Review of feedback: End of October
- Final edit and layout: Early November
- Project steering committee final proofread: Late November
- AI Working Group and Board of Directors Approval: Early December
- English publication targeted for: Mid December
- Spanish translation to follow
Project Steering Committee
Alex Chacon – Antamina
Erman Koc – Airth.io
John Hinton – Rio Tinto
Johnny Witter – Mine Tech Services
Maria Krantz – Epiroc
Pieter van Schalwyk – XMPro
Rob Johnston – CITIC Pacific Mining
Get Involved
To get involved contact us at: gmgprojects@gmggroup.org
Publication Images
Click image to view. Subject to change.