为此，他提出，应尽早确立金融智能体的法律地位。"主要是明确金融智能体的行为边界，明确金融智 能体与金融客户的法理关系，明确金融机构管理者的决策责任，同时建立金融智能体评估审核制 度。"李礼辉说道。 新京报贝壳财经讯（记者黄鑫宇）当前，人工智能正以前所未有的广度与深度重塑银行业等金融机构发 展格局。 中国银行原行长李礼辉近日在参加由四十人高级金融学院和新金融联盟主办的"数智银行家2025年会暨 新金融联盟九周年庆典——数智同行·共启新程"上表示，金融行业、金融机构的价值取向将影响AI替代 的具体速度和深度，即AI替代发生的时间节点、具体岗位和替代比率。 对于上述观点，李礼辉认为，将取决于（金融）智能体的专业性和可靠性，取决于金融监管对智能体的 评估和审核，同时也取决于劳动就业观念和政策的容纳和许可。 "例如，已经投入应用并逐步升级的智能投资顾问，拥有更大的知识面，更专业的分析能力，更冷静的 情绪判断，可将投资顾问从参差不齐的个人专业水平提升到整齐划一的最佳专业水平，有些金融高管认 为智能投资顾问可能替代60%以上的投资顾问岗位。"李礼辉举例道。 当前，金融场景已成为智能体应用的重要赛道。在2025年，广发银行、 ...

Core Viewpoint - The article discusses the evolution and application of financial AI agents, emphasizing their potential to transform the financial industry by enhancing efficiency and accuracy in various tasks, particularly in high-value, technology-intensive areas rather than low-value, labor-intensive sectors [4][5][9]. Group 1: Evolution of AI Technology - Recent advancements in AI technology can be categorized into three main areas: transitioning from unimodal to multimodal capabilities, evolving from AI assistants to AI agents, and reducing energy consumption through innovative algorithms [5][6]. - The latest AI models can process and generate various types of unstructured data, including text, audio, video, images, and code, thus expanding their applicability across different tasks [5]. - AI agents, particularly financial agents, are designed to perform complex tasks in various scenarios, potentially surpassing traditional productivity levels [5]. Group 2: Application Environment of Financial AI Agents - Financial AI agents are being deployed across banking, insurance, securities, funds, and wealth management sectors, gradually replacing human roles, especially in knowledge-intensive positions [7][9]. - For instance, Baidu's digital credit manager can draft due diligence reports in one hour with over 98% accuracy, significantly reducing the time required for such tasks [9]. - The integration of AI in financial advisory roles could lead to a potential replacement of over 60% of investment advisor positions, indicating a shift in the human resource structure within the financial industry [9]. Group 3: Reliability and Economic Viability - The deployment of financial AI agents necessitates advanced security technologies to mitigate risks such as data poisoning and algorithmic biases, ensuring the integrity and reliability of financial transactions [11][12]. - High reliability, interpretability, and economic efficiency are crucial for the successful implementation of financial AI agents, which must be trusted by clients, markets, and regulators [12]. - The focus should be on creating trustworthy AI models that can handle market analysis, customer segmentation, and investment advisory tasks with minimal errors [12]. Group 4: Data Quality and Sharing - The financial sector is data-intensive, and the current data-sharing environment faces challenges such as administrative fragmentation and insufficient circulation of non-public data [14][15]. - To enhance data quality and availability, there is a need for public data to be shared more openly and for private data to be utilized in a market-oriented manner, ensuring privacy and security [15][16]. - Establishing a comprehensive financial database that integrates various data types and sources is essential for the effective functioning of financial AI agents [16].

金融强国建设与中国金融体系现状 金融强国建设是当前金融政策的重要方向，其内涵十分丰富。我的理解是，一个真正的金融强国应当具 备三个方面的特征：高效、稳健和国际影响力，其中国际影响力尤为关键。 对于中国金融体系的现状，通过多年研究，我们对其特点有了较为清晰的认识。中国金融体系可以概括 为"规模大、管制多、监管弱、银行主导"这四个特点。从总体上看，中国金融业规模已经相当庞大。 中央金融工作会议对中国金融体系问题做出了重要判断："支持实体经济的质量欠佳，金融风险比较 多，同时金融乱象和腐败频发，金融监管能力有待提升"。这是一个相当严肃的评估，表明未来金融监 管政策必然会进行调整，实际上相关调整已经开始。 黄益平 金融监管政策的动态平衡 从金融监管政策或金融政策角度看，动态调整可能是一个永恒的主题。最根本的原因是需要平衡效率和 稳定之间的关系。在不同阶段，不同经济体的金融体系所面临的主要矛盾可能有所不同。我们观察到， 一个国家发生金融危机后，监管政策通常会收紧；当收紧到一定程度影响了效率时，又会适当放松。这 种调整是监管的正常过程。 当前中国面临的困难在于需要同时实现两个任务：一方面改善对实体经济的支持力度，这意味着 ...