At the Global Wealth Management Forum 2026 Shanghai Suhewan Conference held in Shanghai from October 16 to 17, 2026, Huang Danggui, Chairman of Bank of China Wealth Management, attended and participated in the meeting.
Huang Danggui stated that AI's current notable advantages lie in its diligence and efficiency, its ability to operate around the clock, and its capacity to greatly improve the efficiency of various basic tasks. For the wealth management industry, it can basically reach the level of a junior assistant, similar to the L2 assisted driving stage in the automotive industry. However, the most core investment decisions, client communication, and other areas still require humans to serve as the main subjects of decision-making and communication. In the future, as AI capabilities continue to evolve, AI's role in the wealth management industry is expected to further upgrade to L3 or even L4, entering the stage of conditional autonomous driving or even highly autonomous driving.
From the practice of the domestic asset management industry, Huang Danggui pointed out that AI is extending from "generalized deployment" to "scenario-based deepening" under the compliance framework. Specifically, first, relying on localized deployment of open-source models, a number of leading institutions are exploring AI practices while ensuring data security; second, focusing on human-machine collaboration, with AI responsible for automated links such as information retrieval, knowledge extraction, and code generation, while investment personnel focus on logical judgment and decision-making control; third, AI tools in vertical fields are constantly emerging. Around business scenarios such as financial report analysis, public opinion early warning, and valuation models, a batch of financial-specific intelligent tools have been gradually implemented, filling the capability gaps of general large models in the financial field.
From the perspective of wealth management companies, Huang Danggui believes that wealth management companies should actively embrace the AI era while firmly safeguarding the security bottom line. First, firmly uphold the compliance bottom line and strictly control the safety gate of AI applications. AI can currently only serve as an "investment assistant" and cannot yet become a "decision-making subject," because model hallucinations may evolve into investment misjudgments, algorithm convergence may amplify market fluctuations, and black-box decision-making poses entirely new challenges to risk control systems. He stated: "AI innovation must not follow an unconstrained trial-and-error path. It must have boundaries, records, intervenability, and rollback capability. Institutional development must be carried out well, and process standards, human defense, and technical defense must run through the entire process of AI exploration and application. At the same time, frontier research should be conducted to explore how to manage highly intelligent AI." Second, learn from the experience of others and appropriately refer to mature overseas practices. AI promotion requires clear division of rights and responsibilities, a gradual promotion pace, and an independent effect evaluation system. Third, deeply cultivate business scenarios and open up the conversion link for AI-enabled business. In the future, it is necessary to move beyond tool-oriented thinking, promote the orderly embedding of AI from "peripheral assistance" to "core processes," connect various single-point capabilities into a complete business chain, and integrate them into the entire process of investment research, risk control, operations, and customer service.