Japanese firms advance AI applications in construction safety and logistics automation

Tokyu Construction and NTT Sonority are collaborating to digitalize hazard prediction (KY) activities at construction sites using voice and generative AI. The initiative aims to transition from handwritten paper records to a system that automatically records and analyzes verbal instructions and safety discussions. Following successful pilot tests using the ‘SonoVo AI’ solution, the company plans a phased rollout during fiscal 2026, with the goal of full implementation across all sites by fiscal 2027.

Separately, ABEJA and JR East have launched a demonstration project at the Takanawa Gateway City development to automate indoor logistics using physical AI. The project successfully utilized dual-arm robots to transfer items, such as bento boxes, to delivery robots. By using foundation models and imitation learning, the system allows robots to adapt to real-world uncertainties like object positioning and orientation, moving toward a fully automated indoor operation model.