In China, local governments are increasingly making use of DeepSeek’s AI models to enhance administrative management and governance, which demonstrates the growing integration of AI in the public management sphere. The Longgang district in Shenzhen, Guangdong province, has taken the lead as the first district in China to comprehensively implement DeepSeek’s cost – effective and high – performance reasoning model. The goal is to improve the governance of its 4 – million – strong population.
One remarkable application is an AI – powered word – processing assistant. It has significantly cut down the time needed for document drafting and proofreading. According to a report on the Longgang government website, previously, proofreading 1,000 Chinese characters took four to five minutes, but DeepSeek’s model can achieve this in just a few seconds.
DeepSeek’s models are also being utilized in combination with the district’s 230,000 surveillance cameras. This helps to simplify the search for missing persons and analyze citizens’ feedback, enabling the government to quickly identify and address crucial issues.
The adoption of DeepSeek’s AI models by local governments comes at a time when the start – up is attracting widespread attention across the country. It has been regarded as a symbol of China’s ability to resist the US’s attempts to restrain its AI development. Major cloud service providers and China’s three major telecom operators – China Mobile, China Telecom, and China Unicom – are incorporating DeepSeek’s models into their services.
While many local governments have already implemented “smart city” solutions using AI, DeepSeek’s models are elevating these applications to a new level. For instance, the government of Kunshan, Jiangsu province, recently announced that it had integrated DeepSeek models into its e – government system, allowing various departments to explore local applications. The police are using DeepSeek models to sift through large amounts of data for clues, and the transport department is implementing them to better predict traffic flows.
