21世纪经济报道记者李览青 "金融大模型将走向专业化、轻量化、合规化，大模型不是量化交易的终点，而将成为智能金融新纪元的起点。" 11月22日，在由南方财经全媒体集团指导、21世纪经济报道主办的"第二十届21世纪金融年会"上，中国科学院大学教授、国家 计算机网络入侵防范中心主任张玉清在主旨演讲中表示。 张玉清指出，当下大模型正在重构量化交易的范式，多智能体与多模态融合将推动量化交易技术不断演进，在未来的某一天， 可能完全靠机器就能在风控强化的背景下抓取信号与重要投资机会。 量化交易瓶颈待突破 张玉清指出，近年来，与基本面基金相比，量化基金无论在业绩表现还是风险控制上的表现都相对突出。 据2023年统计数据显示，全球量化交易在美股市场占比已超60%，在A股市场占比约为20%-30%，2019年到2022年间，A股量化 基金数量翻倍，占主动型公募基金的18%。 然而，量化交易仍存在部分痛点。 张玉清指出，首先是交易策略的同质化。今年8月在A股市场的极端分化行情中，超过七成的量化多头产品跑输了基准指数。张 玉清指出，这主要是由于多数量化策略极度分散，当市场资金集中追捧少数科技权重股时，其收益便被摊薄，难以跟上指数涨 ...

Core Insights - The People's Bank of China has announced the winners of the "2024 Financial Technology Development Award," with Financial One Account winning for the fourth consecutive year [1][2] Group 1: Award Recognition - Financial One Account, in collaboration with Ping An Life and Ping An Technology, received the second prize for the project "Data-Driven and Human-Machine Collaborative Comprehensive Digital Transformation in Life Insurance" [1] - The project aims to create an intelligent operational system covering customers, agents, and internal staff, driven by AI technology [1] Group 2: Innovation in Insurance - The project "New Energy 'Technology + Insurance' New Model Construction," developed by Ping An Property & Casualty and Financial One Account, focuses on the risk characteristics of the new energy industry [2] - It establishes a collaborative insurance ecosystem among car manufacturers, insurance companies, and car owners, promoting green financial innovation [2] - The project features a dynamic pricing model that integrates insurance companies, car manufacturers, and industry stakeholders, enabling automated accident responsibility determination and claims processing [2] - The integrated claims model reduces customer claim waiting time by 60% and decreases secondary accident occurrence by 26% [2] - An AI-based accident simulation platform supports product iteration for car manufacturers, enhancing risk reduction and industry collaboration [2] - The project also includes the development of a global risk map and quantitative model to provide safety and risk management support for new energy enterprises expanding internationally [2] Group 3: Future Outlook - Financial One Account plans to continue its role as a technology output window, collaborating with Ping An's ecosystem and industry partners to explore new intelligent financial models [2] - The company aims to contribute to the service of the real economy, foster new productive forces, and enhance financial risk prevention capabilities [2]

Core Insights - The "Digital Risk Control Project" won the second prize in the national finals of the 2025 "Data Element ×" competition, showcasing a significant achievement among 22,000 participating projects [1] - The competition aims to promote the marketization of data elements and the deep integration of data with industries, featuring 13 industry tracks including financial services [1] Group 1: Project Overview - The "Digital Risk Control Project" addresses industry pain points such as data integration, circulation, and application difficulties, establishing the first "data-risk-ecosystem" digital risk control system in the insurance sector [2] - The project utilizes a robust data foundation and distributed computing capabilities from Ping An Group, creating a comprehensive database covering ten high-quality data categories, with a total data volume exceeding PB level [2] - It integrates over 370 authoritative data sources, forming the first compliance data fusion model and claims knowledge engineering system in the industry, achieving a data standard at DCMM level five [2] Group 2: Technological Innovations - The project has developed a large model and knowledge engineering system for the insurance domain, utilizing trillions of insurance corpus and hundreds of millions of claims data to create an interpretable knowledge graph and intelligent reasoning chain [2] - The automation rate of knowledge has reached 70%, while the data knowledge rate stands at 50%, significantly enhancing risk identification accuracy and control efficiency [2] - The project has empowered over 20 insurance institutions through an inclusive financial open platform, generating economic and social benefits exceeding 10 billion [2] Group 3: Company Strengths - Ping An Group's technology innovation and ecological collaboration are highlighted by this award, reflecting its systemic strength in driving intelligent financial development through data elements [3] - The company has accumulated over 30 trillion bytes of data, covering nearly 250 million individual customers, and has trained large models based on vast data resources [3] - AI has been fully integrated into Ping An's core business, with 89% of car insurance policies being issued in an average of one minute, and the automation rate for personal injury claims reaching 63% [3] Group 4: Future Directions - Financial One Account will continue to act as a technology output window, collaborating with the Ping An ecosystem and the industry to explore new intelligent financial models driven by data elements [4] - The aim is to contribute to the high-quality development of the financial industry, support the real economy, enhance financial risk prevention capabilities, and promote new productive forces [4]

Core Insights - The "Digital Risk Control Project" won the second prize in the national finals of the 2025 "Data Element ×" competition, showcasing a significant achievement for the project among 22,000 entries nationwide [1] - The competition aims to promote the marketization of data elements and the deep integration of data with industries, featuring 13 industry tracks including financial services [1] Group 1: Project Overview - The "Digital Risk Control Project" addresses key industry pain points such as data integration, circulation, and application difficulties, establishing the first "data-risk-ecosystem" digital risk control system in the industry [2] - The project leverages Ping An Group's robust data foundation and distributed computing capabilities, creating a comprehensive database covering ten high-quality data categories, with total data volume exceeding PB level [2] Group 2: Technological Capabilities - The project integrates over 370 authoritative data sources, forming the industry's first compliant data fusion model and claims knowledge engineering system, achieving a data standard at DCMM level five [2] - Ping An has accumulated over 30 trillion bytes of data, covering nearly 250 million individual customers, and has trained large models based on massive datasets [2] Group 3: AI Integration and Impact - AI has been fully integrated into Ping An's core business, with 89% of car insurance policies being issued in an average of one minute, and the automation rate for personal injury claims reaching 63% [2] - In the first three quarters of 2025, AI service volume exceeded 1.292 billion instances, covering 80% of the group's total customer service volume, and AI-assisted sales amounted to 99.074 billion yuan, enhancing customer experience and operational efficiency [2] Group 4: Future Directions - Financial One Account will continue to act as a technology output window, collaborating with the Ping An ecosystem and the industry to explore new models of intelligent finance driven by data elements, contributing to high-quality development in the financial sector [3]

Core Viewpoint - The 2025 China International Service Trade Fair, themed "Digital Intelligence Leading, Service Trade Renewed," showcases advancements in financial technology, particularly by Qifu Technology, under the national "Artificial Intelligence+" strategy [1][4]. Group 1: Company Developments - Qifu Technology has aligned its technological advancements with national policies, focusing on integrating AI into financial services, aiming for over 70% application penetration by 2027 [1][4]. - The company has invested nearly 10 billion in R&D, with over half of its nearly 1,000 patents related to AI, and has a research team exceeding 1,000 members [4]. - Qifu Technology has developed an AI Approval Officer that automates the loan application review process, achieving T+0 approval times and significantly enhancing efficiency while ensuring compliance with regulatory requirements [4][6]. Group 2: Industry Trends - The financial sector is viewed as a prime testing ground for AI technologies, with Qifu Technology aiming to transform AI from a tool into a productive factor through intelligent applications [4][6]. - The company plans to enable one-third of its core business demands to be met through intelligent agents by the end of 2025, indicating a shift towards deeper integration of AI in financial operations [6]. - The evolution of financial intelligent agents is moving from auxiliary roles to core decision-making functions, reflecting a broader trend in the industry towards enhanced efficiency and safety in financial services [6].

Core Viewpoint - The core viewpoint of the report is that the key to AI application in finance is not to engage in a technology race for the sake of AI, but to return to the essence of technology serving business, using ROI as a benchmark to calibrate application paradigms and optimize implementation paths [1][4]. Group 1: Current State of AI in Finance - A productivity revolution driven by large models is quietly occurring in leading financial institutions, indicating a paradigm shift in the industry [1]. - By 2025, it is anticipated that the financial industry will deeply integrate AI and realize the benefits of large model technologies [1]. Group 2: Transformative Practices - A leading bank has reduced complex credit approval report analysis from hours or days to just 3 minutes, with accuracy improved by over 15% [3]. - A top brokerage firm has implemented AI agents to monitor over 5,000 listed companies 24/7, significantly enhancing research coverage and response speed [3]. - An overseas top investment bank has deployed hundreds of AI programmers, with plans to increase this number to thousands, aiming to boost engineer productivity by three to four times [3]. Group 3: Strategic Framework - The report aims to provide a strategic compass that is both forward-looking and actionable, emphasizing the importance of understanding opportunities and challenges, making proactive layouts, and building systematic capabilities [4][8]. - The financial industry is seen as the core battlefield for the comprehensive reconstruction driven by AI, where technology and human wisdom will collaborate to explore the essence of financial services [6][8]. Group 4: Trends and Challenges - The report identifies six core trends driving industry evolution, aiming to provide a strategic roadmap for financial decision-makers and innovators [9]. - The evolution of large models is characterized by a shift from capability exploration to efficiency revolution, with a focus on high-value data rather than just large-scale data [11]. - Financial institutions are moving from experimental phases to large-scale deployment of AI applications, with banks leading the way [12]. Group 5: Implementation Challenges - The implementation of large models in finance reflects the deepening contradictions of digital transformation, requiring institutions to balance fragmented construction, resource allocation, and compliance with safety [14][15]. - Key challenges include data fragmentation, unclear strategic planning and ROI, low tolerance for error in technology adaptation, and lagging organizational talent upgrades [15]. Group 6: Future Outlook - AI is driving financial services towards unprecedented levels of inclusivity, intelligence, and personalization, redefining operational and management models [16]. - The integration of AI with human expertise is expected to accelerate the demand for innovative financial talent, with high-quality private data becoming a core competitive advantage for institutions [16].

Core Insights - Zhongyou Consumer Finance Co., Ltd. was recognized as one of the "Top Ten Smart Finance Cases" for its innovative research and application in intelligent risk anti-fraud [1][2] - The company developed a three-pronged intelligent defense system against fraud, focusing on AI technology misuse [1] Group 1: Technological Innovations - The intelligent defense system includes dynamic biometric deep defense, which integrates multiple live detection technologies and video 3D dynamic analysis to intercept identity impersonation and deepfake fraud attacks [1] - The system employs an innovative graph neural network model (NEI-GraphSage) for precise identification of professional fraud rings [1] - A unified cross-modal decision framework was established to break down data silos, enabling deep integration and collaborative analysis of multi-dimensional information such as images, videos, sounds, and texts [1] Group 2: Application and Impact - The system covers the entire process of loan approval, from pre-loan access to post-loan management, achieving breakthroughs in key scenarios such as marketing fraud detection and identity verification [2] - In 2024, the system successfully intercepted over 1.09 million non-compliant applications and blocked 8,479 fraud cases, recovering marketing losses of 4.55 million yuan [2] - The company aims to deepen the application of this system and continue advancing research in intelligent anti-fraud technologies to enhance its risk control framework [2]

Core Insights - The article highlights the launch of an AI-native mobile banking solution by Kelong Software in collaboration with Henan Kunlun Technology, showcasing a revolutionary "multi-agent collaborative architecture" aimed at providing highly personalized financial services to customers [1][5][6] Group 1: Technology and Innovation - The AI-native mobile banking solution is not merely an enhancement of traditional mobile banking but is built on Kunlun Technology's robust computing power, creating a matrix of specialized financial agents [1][5] - The system utilizes multi-modal interaction and deep learning to accurately understand both explicit and implicit customer needs, akin to having "mind-reading" capabilities [2][5] - The solution allows for real-time analysis of customer profiles, market dynamics, and risk preferences, enabling the dynamic generation of personalized financial, credit, insurance, or investment recommendations [2][3] Group 2: User Experience and Service Transformation - The high-efficiency computing and rapid communication between agents result in a seamless experience from demand insight to solution presentation and transaction execution, significantly enhancing service efficiency and customer satisfaction [3][5] - The AI-native mobile banking solution is designed to integrate seamlessly with existing mobile banking systems and serves as a core engine for intelligent upgrades, providing unprecedented personalized service experiences for retail customers, wealth management clients, and small business owners [5][6] - The launch signifies a transformation in financial service paradigms, emphasizing a service ecosystem that "understands, serves, and empowers" users, thus redefining the depth and breadth of interactions between customers and banks [5][6] Group 3: Industry Implications - The introduction of this solution not only demonstrates the powerful application of domestic core technology in the fintech sector but also outlines a blueprint for future smart financial services, emphasizing the "thoughts become reality" experience [6] - This development marks a significant step in the financial industry's digital transformation, enhancing the quality and efficiency of services to the real economy, and indicates the beginning of a new chapter in fintech driven by artificial intelligence [6]

Group 1 - The article emphasizes the importance of innovative wealth management solutions that provide market-leading data and actionable insights for wealth advisors and investors to make confident decisions [3][4]. - It highlights the wealth management workflow solutions that offer insightful digital tools and features to help wealth management advisors access information in real-time and enhance their efficiency [4][5]. - The wealth data solutions are described as flexible and intuitive, providing advanced analytics and a customer-centric experience [5]. Group 2 - The article invites readers to click on "Read More" for additional information about LSEG's wealth management offerings, indicating a focus on expanding knowledge and engagement in the financial digitalization space [6].

Group 1 - The article emphasizes the importance of innovative wealth management solutions that provide market-leading data and actionable insights for wealth advisors and investors to make confident decisions [3][4][5] - It highlights the flexibility and intuitiveness of the wealth management workflow solutions, which enhance the efficiency and capabilities of wealth management advisors by providing real-time information [4][5] - The article invites readers to learn more about LSEG's wealth management offerings and mentions that registration for the event requires approval [6]