Review of the Second AIoT+X Research Seminar: When Machines Ask for Help—Multi-Agent AI, Human Decisions, and the Future of Smart Factories

On 20 May, 2026, the School of Internet of Things successfully held the second session of AIoT+X Research Seminars. Assistant Professor Dr Oswaldo Ornelas delivered a talk titled “When Machines Ask for Help: Multi-Agent AI, Human Decisions, and the Future of Smart Factories” and explored the cutting-edge intersection of artificial intelligence and industrial manufacturing, offering valuable insights into how humans and AI can collaborate in the factories of tomorrow.

Part.1 Beyond Prediction: The Challenge of Decision-Making

Dr Oswaldo Ornelas highlighted that modern smart factories use sensors, data systems, and AI to detect early signs of machine failure. However, prediction alone is insufficient. Factories must decide whether to stop production, continue the current batch, call an engineer, or delay maintenance. These decisions require more than just alerts—they demand coordinated judgement.

Part.2 The ORCHESTRA Framework: A Human-Centred Multi-Agent AI System

To address this challenge, Dr Oswaldo Ornelas introduced ORCHESTRA, a human-centred multi-agent AI framework for predictive maintenance decision support. The framework brings together multiple intelligent agents: machine monitoring, AI prediction, human review, and maintenance scheduling. AI estimates maintenance urgency, while human operators remain involved in high-uncertainty cases. The goal is not to replace human judgement but to support better, faster, and more responsible decision-making.

Part.3 From Welding to the Factory of the Future

Using resistance spot welding as an example, Dr Oswaldo Ornelas demonstrated how technologies such as XGBoost and PPO can predict maintenance needs and recommend optimal timing. He showed how smart factories can move from simple machine monitoring to coordinated decision support. The key message was clear: the factory of the future will succeed not only because machines are intelligent, but because machines, AI, and people can work together.

Looking Ahead

ORCHESTRA represents an important step toward reliable, human-centred, and intelligent manufacturing systems—where machines detect problems, AI supports decisions, and humans guide responsible action. As the AIoT+X Research Seminar Series continues, the School of Internet of Things looks forward to hosting more sessions that bridge cutting-edge research and real-world industrial needs. Stay tuned for the third seminar!