An article authored by Nexgen Japan was published on February 12 as a World Economic Forum (WEF) Agenda piece. Titled "Why Governance Is the New Infrastructure for Physical AI," the article presents international policy recommendations on AI's expansion into the physical world and the governance frameworks needed to support it.
From "On-Screen AI" to "Physical AI"
In 2025, AI stepped beyond the screen and into our physical economy. AI systems now operate as robots and equipment on manufacturing floors, in logistics hubs, and on construction sites — introducing risks fundamentally different from those in digital spaces.
In digital spaces, a bug can be fixed with a patch. In the physical world, an error becomes an accident. How do we manage irreversible risks? This is the defining challenge of the Physical AI era.
Key Takeaways
- Artificial intelligence is transitioning from an on-screen tool to a physical system operating within the real economy
- Demographic shifts and labour shortages are accelerating the adoption of AI to sustain industrial operations
- Robust governance frameworks are essential to ensure safety in physical environments and to manage operational risks appropriately
The "Survival Strategy for the Physical AI Era" That Executives Must Understand
As labour shortages intensify, leveraging AI effectively requires governance capabilities that surpass mere technical prowess. This paper systematises the decision-making and control structures across three layers of governance: executive, system and frontline.
The frontier of AI deployment has shifted from the office to the field. What are the urgent challenges that executives, those who design AI systems, and frontline leaders must address now? The WEF article presents the full picture.
Read the WEF article (Japanese edition)
Questions we often receive from practitioners
What is physical AI?
Physical AI is AI that works in the real economy, operating as robots and equipment on the front lines of manufacturing, logistics and construction rather than on a screen. In 2025, AI began to move off the screen and into the physical economy. Driven by demographic change and labour shortages, adoption is accelerating as a way to keep industries running. Unlike AI that makes office work more efficient, it is a physical system and bears directly on safety on the front line.
How do the risks of physical AI differ from those of on-screen AI?
Failures cannot be undone. An error that a bug fix would resolve in the digital world becomes an accident in the physical world. Ensuring the safety of physical AI and managing its operational risks therefore requires a governance framework, not technical capability alone. Nexgen Japan proposed to the World Economic Forum that governance should be positioned as the new infrastructure for physical AI.
Who should be responsible for governing physical AI?
Responsibility is shared across three layers: executive, system and frontline governance. Executives set the intent, accountability and non-negotiables, including what cannot be delegated to an algorithm. System governance builds that intent into design by deciding which decisions are automated, which are augmented and where human approval is mandatory, and by defining stop rules and safe states. Frontline workers need the authority to override AI, the judgement to interpret confidence and constraints, and the right to intervene. Nexgen Japan set out this framework through the World Economic Forum.
Related Seminar
The Physical AI Shock — Robots Reshaping Industry and Work
From "on-screen AI" to AI operating in the field — taking the governance questions raised here into practice. This programme systematises global trends in Physical AI, the three-year outlook, and sector-specific impacts on manufacturing, logistics, healthcare, and construction, with a workshop to design your organisation's AI collaboration blueprint.
View seminar details →