As 2025 marks the conclusion of China's 14th Five-Year Plan, a retrospective look reveals Changsha's economy demonstrating remarkable resilience with increasingly robust momentum for high-quality development. To provide in-depth analysis of Changsha's economic achievements, Changsha Evening News launches its "2025 Changsha Year-End Economic Review" series, focusing on industrial upgrading, consumption expansion, employment security, and export quality improvement to decipher development logic and document progress.
What was the hottest keyword of 2025? "Artificial Intelligence" undoubtedly takes the crown. From versatile humanoid robots to sophisticated industrial robotic arms, AI is rapidly integrating into daily life. As the final mile of technological implementation, expanding application scenarios proves particularly crucial.
The recent official release of the "Changsha 'AI+' Action Plan" (Chang Zheng Fa [2025] No. 13) charts a blueprint for Changsha's AI industry development, placing application scenarios at the forefront. "Changsha boasts rich AI application scenarios and is actively expanding their scale," noted a relevant official from the Changsha Bureau of Industry and Information Technology. The Action Plan covers the entire chain from core technology R&D to industrial cluster development, scenario application promotion, foundational support, and talent policy guarantees, forming a complete innovation-to-industrialization ecosystem while driving technological innovation and industrial upgrading through open scenarios.
With the clarion call sounded and the ambitious goal of building an "AI Scenario Application Demonstration Zone" anchored, what are Changsha's key initiatives? Examining the simultaneously released first batch of Changsha's "AI+" Demonstration Scenarios List reveals 14 blossoming scenarios across industrial development, consumption upgrading, public welfare, and governance capabilities. Recently, reporters visited representative enterprises to explore how "AI+" creates a diverse application ecosystem.
Using wheels as the medium to transform computing power into productivity "Delivery in progress, please note!" Upon arriving at Changsha Xing Shen Intelligent Technology Co., Ltd., an autonomously driving unmanned delivery vehicle immediately captured attention. Moving like a seasoned driver, it nimbly navigated obstacles without guidance, precisely parking at its destination with skilled operations that drew admiring glances from pedestrians.
Xing Shen Intelligent not only secured a spot on Changsha's first batch of key AI enterprises in robotics manufacturing but also saw its "Unmanned Delivery Based on Fusion Environmental Perception Large Models" project selected for the first "AI+" Demonstration Scenarios List in the consumption upgrading category, directly impacting daily parcel delivery experiences.
"Last-mile delivery represents the costliest, least efficient, and most challenging segment in the logistics chain, accounting for over 30% of total logistics costs," explained a company representative. Born from China's earliest autonomous driving research team at National University of Defense Technology, the enterprise leverages over 30 years of technical accumulation to reconstruct urban logistics' "capillaries" through AI, liberating humans from repetitive, strenuous, and hazardous final-mile transportation.
Currently, its unmanned vehicles achieve "L4-Highly Automated Driving" intelligence, capable of intervention-free autonomous operation within designated areas. Where does this "intelligence" and "autonomy" originate?
On one hand, vehicles deploy multi-source, multi-level fusion environmental perception large models based on deep learning. Using vision as the core while integrating millimeter-wave radar, cameras, and lidar data, they possess "superhuman" sensory systems with environmental comprehension capabilities. "In extreme weather like rain, snow, or fog where cameras may falter, radar remains effective; at complex urban village intersections, large models accurately identify traffic participants and predict their trajectories," the representative elaborated.
On the other hand, fully domestically-developed computing infrastructure ensures technological autonomy. Choosing this path addresses not only supply chain security but also demonstrates Chinese manufacturing prowess. "It holds inherent cost advantages in supply chain coordination and mass production," the representative emphasized, noting that high computing power doesn't necessarily mean high costs, as deep algorithm optimization and performance enhancements prevent "using a sledgehammer to crack a nut" computational waste.
Today, Xing Shen Intelligent offers diverse vehicle models: unmanned logistics vehicles with large compartment delivery, temperature-controlled meal delivery carts, automated retail vehicles, and "kangaroo" vehicles carrying small robots for doorstep delivery. These operate across Hunan Xiangjiang New Area, Changsha County, Changsha Economic Development Zone, Central South University, and National University of Defense Technology.
Vehicle dispatch proves exceptionally convenient, upgrading from "vehicles waiting for goods" to "goods finding vehicles." Users simply confirm pickup and delivery points via mini-programs, with information transmitted to a digital platform for shared transport capacity management where systems autonomously coordinate unified dispatch. Data shows this nationally pioneering regional shared unmanned transport scheduling system improves efficiency by approximately 40% compared to traditional manual delivery, with end-point collaborative distribution boosting transport utilization by over 50%.
As a local enterprise, Xing Shen Intelligent received multifaceted government support during critical development stages, obtaining specific testing zones and licenses through pilot mechanisms while establishing data interoperability with users. "The 'Changsha Model' of 'government guidance, enterprise leadership, and scenario traction' is key to our commercial closed-loop success," the representative noted, viewing the company as a "connector" in Changsha's AI ecosystem—linking domestic computing power and sensor industries below while connecting e-commerce, retail, and medical applications above. "Through our wheels, we transform AI's computing power into flowing urban productivity."
In Changsha, wheel-based intelligence exploration extends far beyond last-mile delivery. CIDI Technology Co., Ltd., recently listed on the Hong Kong Stock Exchange's main board and also selected for Changsha's first batch of key AI enterprises, is deeply integrating large models into complex smart driving decisions. Its led "Large Model Application Research in Smart Driving" project aims to enhance vehicles' open-road perception, prediction, and planning capabilities through cutting-edge AI technology, providing key support for higher-level autonomous driving and smart transportation systems.
Using AI as the brain to advance from "empirical medicine" to "computable medicine" Artificial intelligence not only optimizes transportation "arteries" but also precisely diagnoses human health.
How should prescriptions address patients with multiple complications? While traditional diagnosis heavily relies on "veteran doctors" experience, AI offers new approaches: doctors input primary symptoms and diagnoses into systems that recommend multiple suitable formula combinations within seconds based on knowledge graph analysis. Physicians can review detailed explanations for each recommendation, determining final prescriptions combining clinical experience with system suggestions.
This intelligent "AI assistant" is the traditional Chinese medicine multimodal large model—"Simiao AI"—developed by Hunan Boji Life Technology Co., Ltd. The company earned selection for Changsha's first "AI+" Demonstration Scenarios List in public welfare through its "TCM Compound Optimization and Recommendation Based on Multimodal Large Models" project.
Boji Life founder and general manager Feng Wu noted that TCM represents a data treasure trove, but its knowledge system long depended on experiential transmission lacking structured, standardized analytical tools. Thus, AI technology's rapid development arrived opportunely, "allowing us to interpret TCM compounds from molecular levels, achieving the leap from 'empirical medicine' to 'computable medicine'."
"Simiao AI" embodies this vision's practical realization. Through Boji Life's self-developed optimized Transformer architecture, it integrates "formula-herb-molecule-target" four-layer data, advancing TCM research from single-dimensional to fused multidimensional analysis. "When doctors input formula combinations, Simiao AI not only predicts effective treatments but traces which herb components act and their bodily targets," Feng highlighted, identifying this full-chain, explainable analytical capability as its core competitive distinction from other TCM AI systems.
Dramatically reduced literature review time, improved solution comparison efficiency, and decreased repetitive labor—from requiring 25-30 minutes per complex case to cutting clinical expert decision-making time by 50%—demonstrate Simiao AI's immediate impact. Feng reported that a provincial TCM hospital assisted over 3,000 outpatient cases during a six-month trial, achieving 15% reduced patient revisit rates and significantly improved first-visit prescription effectiveness. In primary healthcare, a community health center pilot saw resident satisfaction with TCM services rise 20 percentage points within six months.
As a project signed during Changsha's global R&D center city inauguration, Boji Life's growth benefits from local innovation ecosystems. Feng indicated their practice offers insights for local AI industry development: leveraging regional advantages for deep innovation in vertical segments to solve critical problems.
Currently, Boji Life explores integrating Simiao AI's TCM knowledge base and health analysis capabilities into self-developed multimodal emotional companion robots like "Xiao Bo" and "Xiao Mei." Future integration may enable robots providing emotional companionship while offering personalized wellness advice based on TCM "preventive treatment" and emotional regulation principles, creating亲切, intelligent TCM accessibility pathways in household settings.
AI's medical empowerment extends further. Wisdom Eye Technology Co., Ltd., also selected for Changsha's first batch of key AI enterprises, developed the "Bianshi Foundation Model"—Hunan Province's first approved medical large model registered with the Cyberspace Administration of China's generative AI (large language model)备案—providing essential foundational cognitive support for clinical assistance and imaging analysis scenarios.
From unmanned street deliveries to TCM-assisted diagnosis, these vivid practices epitomize Changsha's scenario-driven AI industry development. How to advance further from current achievements? The Action Plan establishes six key projects, including the "Application Scenario Promotion Project" that will build scenario service platforms, open industry scenario resources, conduct scenario investment and supply-demand matching, and publish scenario requirement lists. With ongoing policy propulsion, a symbiotic "industrial rainforest" accelerates growth, destined to make Changsha's AI application ecosystem flourish with luxuriant branches and leaves.
