November 2026 Summaries

Superb AI presented its “Vision Intelligence for Public Safety” approach at the 2026 Korea-Singapore Safer World Innovations Programme seminar, describing how patrol robots, existing CCTV systems, drones, and 3D digital twins can expand surveillance coverage and centralize monitoring. Its platform reconstructs environments from ordinary video rather than requiring CAD, BIM, or LiDAR data, enabling operators to locate robots, view feeds, dispatch devices, and receive mapped alerts for hazards such as fires, collapsed individuals, loitering, traffic violations, and violence. The company’s ZERO vision foundation model supports plain-language detection requests, explains the basis for alerts, searches recorded footage across cameras, and incorporates operator decisions to improve performance at individual sites while retaining human oversight of emergency responses. Examples included Gimcheon City’s operational safety-monitoring system, which reportedly achieved 90% accuracy for key objects in CCTV footage, and a proof of concept at Incheon International Airport for passenger-flow analysis, wait-time prediction, and cross-camera tracking. Superb AI emphasized that the technology can operate in cloud, on-premises, or edge environments, including closed networks, and can begin with agencies’ existing cameras and a targeted use case.