In a context where the world is seeking solutions to deploy innovation at scale responsibly, Mr. Eisaku Ito, President and CEO of Japan’s leading heavy industry group, Mitsubishi Heavy Industries (MHI), shared profound insights at the World Economic Forum (WEF) on how to “wield AI with responsible hands.”
The Distinction of AI in the Manufacturing Sector
Artificial intelligence is reshaping how industries operate around the world. Yet most of today’s discussion revolves around large language models (LLMs) and their impact on office work.
For manufacturers, the outlook is different. Although the potential for productivity gains is huge, applying AI in industrial manufacturing raises hard questions: how can we fully trust algorithms when they control critical infrastructure on which human lives depend? How safe is it to share proprietary knowledge with partners in the supply chain?
At Mitsubishi Heavy Industries (MHI), we see solving these challenges as a priority, not only for the benefit of the business but also for the sustainable development of society.

Immense Potential for Productivity
In fact, AI is not entirely new to advanced manufacturers. If AI is defined as collecting big data and using algorithms to find insights, many companies have been doing this for years to optimise production and help customers operate equipment.
In MHI’s factories, AI is used to automate processes, make robots and machine tools smarter, and ensure consistency in quality inspection.
There is still great scope to go further, for example in welding, a core process in assembling gas turbines, ships and nuclear reactors. This is a highly refined skill, in which experienced welders rely on a “sixth sense” to produce a perfect weld. We are digitising this process by recording welders’ movements to train AI. The aim is to combine human skill with machine consistency, while moving workers out of hazardous environments into safer roles.
We are also developing algorithms for operational technology (OT) systems. Every gas turbine we sell has built-in capabilities to detect the smallest anomalies, allowing maintenance to be predicted weeks before a failure occurs.
Challenges in Security and Data Reliability
However powerful it is, AI is ultimately a tool. The competitive advantage it provides can erode quickly if companies do not know how to combine it with human experience and skill to create solution packages that are hard to replicate.
A bigger concern is the reliability of algorithms. AI models trained on third-party data sometimes do not produce accurate or reproducible results. The best results usually come from training on in-house data, but the challenge is whether there is enough of it.
The security of proprietary data is also a major barrier to building smart supply chains. Although investment in hyperscale data centres is rising, many manufacturers still want tight control over their data. This could drive demand for edge data centres located on factory sites, a promising market MHI is targeting.
Maintaining the “Monozukuri” Spirit and Human Skills
The biggest concern for manufacturing today is that the AI boom is drawing resources away from the development of physical production technologies. More and more students are pursuing software and algorithms, while courses in hardware and craftsmanship, or Japan’s spirit of “Monozukuri” (the art of making things perfectly), are short of learners.
AI has the potential to transform manufacturing, but it should be seen as a tool that supports rather than fully replaces the human role. For AI to have a truly positive impact, it must be applied by thoughtful, responsible hands that respect the value of craftsmanship.
Mr. Eisaku Ito President & CEO, Mitsubishi Heavy Industries (MHI), Japan

