JKMRC Friday Seminar 2026: Why Autonomous Mining Systems Struggle with Decision-Making

Item: JKMRC Friday Seminar 2026: Why Autonomous Mining Systems Struggle with Decision-Making
Rating: 4
Author: PostSingularity Institute

by Mohsen Yahyaei

Applied Sciences & Engineering Engineering & Technology TTOther technologies and applied sciences TTUMining technology and engineering

🎙 Mohsen Yahyaei 👥 6K 📅 March 8, 2026 ⏱ 67 min 👁 181 📄 expert opinion 🧭 2026-08-16

Available in: English (current) Français

Keywords

autonomous miningdecision-makingprocess automationtrade-offshuman-in-the-loop

Summary

In this seminar, Professor Mohsen Yahyaei, Director of the Julius Kruttschnitt Mineral Research Centre (JKMRC), argues that autonomous mining systems struggle not because of weak control models or immature AI, but because the critical decision-making layer—where competing objectives are negotiated—remains informal, fragmented, and largely invisible. He contrasts the ‘visible operation’ (physical processes, sensors, automation) with the ‘hidden operation’ (human negotiations, trade-offs, context). Current automation architectures, including advanced process control (APC), model predictive control (MPC), and AI-based optimizers, assume systems are predictable, measurable, and optimizable, but real operations face conflicting objectives (throughput vs. recovery, energy vs. productivity), changing constraints, and incomplete data. He critiques the trend of removing humans from control rooms, arguing that AI cannot capture the contextual knowledge that human operators possess. To address this, he proposes Communicative Process Automation (CPA), a hybrid approach that combines traditional system thinking for unit-level optimization with complex system thinking for plant-wide decision-making. CPA introduces a ‘Decision Systems layer’ with specialized agents that negotiate trade-offs, make rationale explicit, integrate human expertise, and adapt under uncertainty. The agents are not process experts but experts in negotiation, and they are trained on general decision-making concepts, not proprietary plant data. The system includes a knowledge center to store decisions and outcomes, enabling scalability and consistency. Yahyaei illustrates the concept with a mineral processing example (grinding, classification, flotation) and emphasizes that true autonomy requires making the hidden operation visible, measurable, and intelligently supported.

236 words

Critical Evaluation

Value of the Information & Strength of the Argument

The presentation offers valuable insights by identifying a critical gap in current automation: the lack of a formal system for managing trade-offs and decision-making under uncertainty. The argument is well-structured, moving from the limitations of existing approaches (optimization, AI) to the proposal of a new framework (CPA). The speaker uses practical examples from mineral processing to illustrate the challenges, making the argument concrete. However, the proposal is conceptual and lacks empirical evidence or case studies demonstrating its effectiveness. The argumentation is persuasive but relies on the speaker’s authority and experience rather than rigorous data.

Scientific Rigor, Source Quality, Title Accuracy

The speaker is a recognized expert (Director of JKMRC) and the content is based on practical experience and a proposed framework. However, it is an opinion/expert talk without peer-reviewed data or empirical validation. The title accurately reflects the content, which focuses on the challenges of decision-making in autonomous mining systems. The talk does not cite specific external sources, but it references industry efforts (ABB, Honeywell) and academic concepts. The description provides links to the SMI webinar archive and YouTube channel, which are relevant for further exploration.

195 words

Title / Content Match

The title accurately reflects the content, which focuses on the challenges of decision-making in autonomous mining systems.

Quality & Reliability

8/10

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

SMI Webinars — Link to past webinars from the Sustainable Minerals Institute, where this seminar series is hosted.
SMI YouTube Channel — YouTube channel for the Sustainable Minerals Institute, hosting previous seminars.

Concurring Sources

SMI Webinars — The seminar series is part of the SMI's ongoing efforts to discuss topics in the minerals sector.

Contribution & Novelties

The presentation offers a novel perspective by shifting the focus from optimizing control loops to formalizing the decision-making process itself. It introduces Communicative Process Automation (CPA) as a new architectural layer that structures negotiation between competing objectives, making trade-offs explicit and integrating human expertise. This approach contrasts with current AI and optimization methods that assume a known optimal solution. The proposal is original and could inspire new research directions in autonomous systems.

Pour aller plus loin :

Complex systems theory — Relevant for understanding the shift from reductionist to holistic approaches.
Model predictive control — Background on a key control technique mentioned in the talk.
Human-in-the-loop — Concept related to integrating human expertise in automated systems.

116 words

Radar Profile

The radar profile shows high scores in information quantity, quality, and technical level, with a slightly lower score in global reliability due to the lack of empirical validation. This indicates a well-informed and technically detailed presentation, but one that is primarily based on expert opinion rather than peer-reviewed evidence.

Reliability 8/10