The 2026 Policy Address proposes to examine and set up an “artificial intelligence city brain” integrated management system. Through big data analytics, it will enable the Hong Kong Special Administrative Region’s government departments to access real-time information and strengthen their capacity for day-to-day city management and handling of emergencies. The goal is to complete the first stage of construction in the first quarter of 2028, at an estimated cost of around HK$200 million ($25.48 million). Within the constraints of limited time and budget, the most practical approach is not to start from scratch, but to integrate and upgrade existing digital infrastructure.
The core of the “AI city brain” is not technological showmanship, but solving the inconveniences that residents face every day. When a typhoon hits, which streets are likely to flood? When the MTR breaks down, how can passenger flow be swiftly redirected? When a water main bursts, can immediate warnings be issued? These are the issues that residents genuinely care about.
In fact, Hong Kong already has a number of “fragmented” systems developed independently by various departments. For example, the Drainage Services Department has developed the Mosaic Model Map, which combines hydraulic models with data from the Hong Kong Observatory to identify areas with a higher risk of flooding. This allows the department to install flood barriers in advance and dispatch drainage machinery — along with drainage robots — to stand by. During Super Typhoon Ragasa last year, the Emergency Monitoring and Support Centre operated continuously for more than 70 hours, and all flooding cases were handled within one to two hours. These systems have taken the first step toward a people-oriented direction, but they operate in silos; what is lacking is a unified platform that can link data together.
City management often fails not because data is absent but because data cannot be aligned on the same map. At present, one department may manage road infrastructure while another handles population and public services — each running its own operations, but unable to coordinate around the same spaces and targets.
Geospatial coordinates are precisely the key. A city is a geospatial entity, and all city-related objects have location attributes; location is the thread that ties together different types of urban data. The experiences of both Hangzhou and Shanghai offer Hong Kong valuable reference points. Since Hangzhou began building its “city brain” in 2016, it has consistently used Geographic Information System (GIS) as the underlying logic for aggregating data. It has gradually integrated systems across the city, including transport, policing, urban management, and healthcare. In just two years, Hangzhou reportedly moved from 5th to 57th in the national ranking of congested cities, and ambulance access times were cut by half. The evidence points to GIS and unified spatial frameworks as the driver — because they allow AI to analyze across domains and support faster decisions rather than isolated, silo-based reporting. Shanghai, meanwhile, consolidated information functions across multiple departments and used “one map” as the foundation for urban governance. These cases demonstrate that GIS is not merely one of many systems; rather, it acts as the “brain’s nervous system” that allows all systems to communicate with one another.
Most importantly, Hong Kong is not starting from zero. We already have GIS-centered systems that provide the foundation for an integrated approach. The Common Spatial Data Infrastructure has more than 1,100 data sets; in 2025, it recorded over 2.05 million downloads and 95 million requests for application interface services. The Underground Utilities Information System consolidates underground utility data held by government and public utilities organizations. And the Common Operational Picture (COP) supports real-time incident information sharing across more than 20 government departments. COP has been developed over more than 10 years and has been operating since 2020, based on GIS. The evidence here is strong: The infrastructure for spatial integration already exists.
The limitation is scope. COP is currently used mainly for disasters such as landslides and flooding. That narrow coverage means the potential value of a shared platform is not fully realized. This is where an “AI city brain” could make a tangible difference — by extending and integrating the existing geospatial foundations beyond current use cases.
Therefore, the next step is straightforward: build an “AI city brain” that is anchored in three elements — people, geospatial data, and the COP operating framework. By integrating the existing GIS-centered systems, and adding supporting technologies such as geospatial AI, digital twins, satellites, drones, and sensors, interdepartmental collaboration can be strengthened through a single platform, improving data interoperability, and raising governance efficiency. All in all, this would enable AI to understand how to provide optimal decisions for urban governance, and help deliver a safer, more resilient “smart city” for the public.
The author is an adjunct professor at the School of Computing and Data Science, Department of Geography, Faculty of Social Sciences, the University of Hong Kong.
The views do not necessarily reflect those of China Daily.
