证监会科技司刘铁斌：构建规范、健康的资本市场人工智能生态体系

Core Viewpoint - The China Securities Regulatory Commission (CSRC) is actively monitoring the application and development trends of artificial intelligence (AI) in the capital market, aiming to enhance the application level of AI while ensuring compliance and innovation [1][2]. Group 1: AI Application in Capital Markets - The capital market has shown a positive trend in exploring AI applications, characterized by leadership from the securities industry, followed by funds and futures, with leading institutions setting examples [1]. - AI applications have expanded to core areas such as investment research, financing services, investor services, internal control management, information technology, and regulatory supervision [1]. Group 2: Challenges in AI Application - Challenges include non-unified data standards, insufficient high-quality data sources, and issues related to large model "hallucinations," which hinder the large-scale and in-depth application of AI [2]. - Regulatory frameworks tailored to the characteristics of AI technology are still maturing, and there is a need for improved precision and foresight in regulation [2]. - Traditional risks such as cybersecurity and data security remain significant, alongside emerging risks like "data poisoning," which require enhanced risk management capabilities [2]. Group 3: Strategic Framework for AI Development - The basic work approach for building a healthy AI ecosystem in the capital market includes balancing innovation promotion with regulatory compliance, respecting technological evolution, and collaborating across industry and institutional ecosystems [2][3]. - Key strategies involve establishing a differentiated regulatory framework, promoting pilot projects for "AI + capital market," and enhancing resource sharing for computing power, data, and models [3]. Group 4: Institutional Ecosystem for AI - Institutions are encouraged to develop governance structures and strategic plans that align AI applications with their development, ensuring compliance with national ethical standards and legal regulations [3]. - Emphasis is placed on improving data governance, using authoritative data sources, and enhancing data quality to support AI applications [3]. - Institutions must establish risk assessment and emergency response mechanisms to monitor high-risk business scenarios continuously [3]. Group 5: Compliance and Risk Management - Compliance with national and industry regulations is critical, including establishing content compliance mechanisms and preventing data security risks [4]. - Strengthening network security management and enhancing model management capabilities are essential for ensuring the safe operation of AI-related information technology systems [4]. - A collaborative mechanism involving human oversight and model assistance should be established to mitigate the inherent limitations of AI [4].