The governance of a city requires a plethora of  complicated systems that in the past have brought with them challenges in planning, coordination and execution. In the age of technology, AI is lending a powerful helping hand to make these systems far more efficient.

DeepSeek, the Chinese AI large language model (LLM) that has gained so much world attention, has now been connected to the e-government service platforms of many cities, including Beijing, to work its magic.

When a transportation alert emerged near an amusement park in Beijing's Changping district, the DeepSeek-connected city operation system quickly calculated the deviation between the transportation capacity and volume.

In less than 20 minutes, the system generated a comprehensive emergency plan. The real data later showed that the efficiency of acceptance for carriage by buses surged by 65 percent, cutting the waiting time of passengers by half an hour.

Before DeepSeek was connected to the system, it would take at least two hours to coordinate transportation capacity as the staff had to conduct cross-over analysis of data from multiple systems in person after an alert was raised, according to a staff member of the Changping municipal commission of urban management.

The local transportation operation data of Changping, emergency plans, laws and regulations have been imported to the system using DeepSeek, forming a localized knowledge base. The LLM can thus conduct deep learning and accurate analysis to provide staff with more intelligent service and more precise localized solutions.

Electric bicycles are the cause of much distress in Longhua district, Shenzhen. Despite the convenience electric bicycles have brought to people's daily transportation, they also cause great challenges to urban governance, such as traffic accidents. With the help of AI, big data and Internet of Things (IoT), Longhua managed to create a closed loop of incident identification, distribution, disposition and evaluation.

Relying on an AI supported platform, Longhua can recognize and capture images of people riding electric bicycles without helmets via monitoring cameras.

When an illegal activity is discovered, it can be automatically detected by the LLM incorporated into the platform. Off-site enforcement can be done via mini programs, short messages and 5G phone calls. Analysis of the overall situation of the activity and proposal of effective suggestions for improvement can be made based via LLM and big data analysis.

In Xindu district, Chengdu, southwest China's Sichuan province, AI, big data and IoT are making a meaningful contribution to the building of a smart city.

Three smart platforms for city operation, data resources and appeals from citizens have been established, integrating over 20 smart applications and 366,000 sensing devices made up of 65 types. With such a smart system in place, 17 types of problem areas can be handled and traced within one single network, taking as little as two hours.