数据信任

Search documents
金融业如何与大模型“共舞”
Jin Rong Shi Bao· 2025-08-19 01:40
Core Insights - The financial industry is undergoing a profound transformation driven by large models, which are reshaping roles, functions, and business models within the sector [1][3] - The application of large models in finance is transitioning from a phase focused on technological validation to one that emphasizes commercial value and systematic integration [3][5] - Data is becoming a critical element in the evolution of large models, with the need to address data fragmentation and enhance data trust and governance [5][6][7] Group 1: Development Trends - The application of large models in finance is moving towards enhancing core revenue-generating areas and evolving from efficiency tools to collaborative decision-making partners [3] - The financial industry is actively embracing large models through two main approaches: training general large models with financial data and developing specialized financial models by AI startups [3][4] Group 2: Challenges - The implementation of large models faces three core challenges: high costs, scarcity of professionals who understand both finance and AI, and difficulties in managing organizational culture and processes [4] - The industry must confront the challenge of data governance, as data is currently seen as the largest obstacle in the application of large models [7] Group 3: Data Utilization - Financial institutions are encouraged to activate dormant data, develop synthetic data, and advance data standards to leverage high-value data resources [5][6] - Trust in data is essential, categorized into three levels: trust in data collection and usage, trust in the data itself, and trust in data creators [6]
全球金融系统为何需要可靠的高质量数据?
Refinitiv路孚特· 2025-03-10 06:00
Core Viewpoint - The article emphasizes the critical importance of high-quality, trustworthy data in the financial industry, especially as AI technologies reshape the landscape and increase demand for data [1][2][4]. Group 1: Data Integrity and Management - The financial industry is witnessing a surge in demand for reliable data, with an annual growth rate of approximately 40% since 2019 [1]. - Ensuring data integrity and relevance is essential for large language models (LLMs), as poor data can lead to unreliable AI outcomes, including hallucinations and biases [2][4]. - The implementation of data transparency, security, and integrity is crucial for compliance and building customer trust, exemplified by the use of "watermark" technology in financial data [5][6]. Group 2: Digital Rights Management - Digital rights management is increasingly important, with clients expecting clear definitions of data source availability, AI policy responsibilities, and effective measures against intellectual property infringement [4][6]. - The establishment of a robust framework for digital rights management is necessary to ensure compliance and responsible AI usage [9]. Group 3: Regulatory Framework and Collaboration - The article highlights the need for standardized definitions of "data trust" across the industry to enhance data quality and facilitate its circulation [6][10]. - Collaboration across the industry and the establishment of interoperable regulatory frameworks are key to fostering high-trust AI systems and promoting innovation in the global financial sector [2][10]. - Regulatory initiatives, such as the EU's AI Act and the NIST's AI Risk Management Framework, provide guidance for safe and fair AI practices in the financial services sector [7][11]. Group 4: Future Outlook - The importance of high-quality data will continue to grow as the industry develops AI tools, necessitating a deep understanding from policymakers [8][11]. - The article calls for a collective effort from financial and tech companies, regulators, and users to navigate the challenges posed by evolving AI technologies [11].