AI+Finance enters a phase of large-scale implementation: how to make it usable, controllable, and traceable
English text is being prepared.
On October 23 at the Shanghai Suhewan Conference, Zhu Jian, Chairman of Guotai Haitong Securities, said that the deep integration of technology and finance is already an inevitable choice of the era, and mentioned that the CSRC recently made clear it will issue and implement opinions on the safe and compliant application of artificial intelligence in the securities and futures industry. Wang Cheng of Tongji University pointed out that financial decision-making has many variables and complex correlations; simply increasing data, computing power, and algorithms may not solve deep-water problems such as macroeconomic forecasting and trading simulation, and finance has high requirements for explainability and anomaly traceability, so new technologies need to integrate with existing methods rather than simply replace them. Attendees believed that the industry needs to adhere to independent and controllable hardware and an open software ecosystem, and promote networked computing power and heterogeneous computing power scheduling.
Body
“科技与金融的深度融合，已是时代的必选项。”10月23日，国泰海通证券董事长朱健在上海苏河湾大会高端对话活动上称。
朱健提到，监管部门持续推动科创板第五套上市标准向人工智能、大模型、量子计算等前沿领域扩展，证监会近期明确将发布实施人工智能在证券期货业安全合规应用的意见。
静安区委副书记、区长翟磊表示，静安作为上海“五个中心”建设重要承载区，集聚各类金融机构800余家，金融业增加值规模位居全市前三，正大力实施“人工智能+”行动，实体化运作全市首个AI场景应用促进中心。
具体到AI金融的发展难点，同济大学计算机科学与技术学院副院长王成称，金融决策变量多、关联复杂、随机性强，单纯依靠更多数据、算力和算法，未必能解决宏观预测、交易推演等“深水区”问题，需要推动多范式融合。
王成强调，金融对可解释性和异常追溯要求高，新技术需与现有方法融合，而非简单替代。
当前，AI与金融的融合已从单点技术验证进入体系化、规模化生产阶段。在圆桌讨论中，与会嘉宾认为，行业需坚持硬件自主可控、软件生态开放，提升网络确定性、安全性和时效性，完善存储底座和智能运维治理，推动算力网络化和异构算力调度，让AI在金融核心场景中“用得上、管得住、可追溯”。