The 2026 Bund Summit, themed "Technology, Geopolitics and the Global Economy," was held in Huangpu District, Shanghai, from October 19 to 22, 2026. Qian Bin, Vice President of Bank of Communications Co.,Ltd. (ASX: 601328), attended and delivered remarks.
Qian Bin stated that AI in the financial industry is moving from pilot exploration to large-scale deployment and has become the core driving force behind the industry's reshaping.
From the perspective of capital investment
AI has become the "main battlefield" for financial institutions' strategic spending. AI investment is no longer a simple "tool procurement" but a deep-level "capability building," involving systematic combined investment in computing power, models, data, platforms, intelligent agents, and business system transformation.
From the perspective of the technology foundation
AI has been upgraded to a core technological capability of financial institutions. AI is no longer a simple upper-layer application but a foundational, core capability platform, increasingly exhibiting characteristics of componentization, modularization, and public infrastructure-ization, becoming a bottom-layer public foundational capability that provides support and services for various business systems. It has been comprehensively embedded into the entire fintech management system, delivering value for business decisions and operations.
From the perspective of application scale
AI development among Chinese financial institutions has entered the stage of "large-scale value realization." In the first half of 2026, banking AI development exhibited three notable characteristics: first, the number of scenarios at state-owned large banks generally reached over 500; second, Token consumption moved from the "million level" to the "tens of billions level"; third, intelligent agents shifted from pilot programs to systematic deployment, entering comprehensive application.
From the perspective of the policy environment
On June 18, 2026, the National Financial Regulatory Administration issued the "Guiding Opinions on the Safe Development and Application of Artificial Intelligence in the Banking and Insurance Sectors," guiding the entire financial industry away from the early stage of free growth and free trial-and-error with an emphasis on development over security, and gradually moving from "usable" toward a new milestone of "manageable, controllable, and traceable."
Qian Bin stated that financial AI has achieved notable results in multiple areas. In risk prevention and credit approval, AI is applied comprehensively before, during, and after lending. Taking Bank of Communications Co.,Ltd. (ASX: 601328) as an example, the bank developed an intelligent agent for trade background review that has streamlined the entire process. The manual steps required for a single forfaiting transaction were reduced from 9 to 4, and transaction time dropped from 2 hours to under 30 minutes. The anti-money laundering alert rate has also improved significantly. In market trading and operational efficiency improvement, AI is widely used for foreign exchange inquiries, money market inquiries, bond counterparty recommendations, and trade report generation; in intensive operations, AI is used for intelligent document recognition and entry, business review, and voucher processing; in technology R&D and operations, large model technology has greatly enhanced research and development capabilities. AI is also increasingly applied in marketing, customer service, and intelligent agent services.
He believes that the future combination of intelligent agents and payments holds enormous space and broad prospects, especially in cross-border payments, anti-money laundering, international balance of payments declarations, exchange rates, and interest rates. In any case, the changes AI brings to the financial industry are fundamental, delivering massive cost reductions and efficiency gains.
Risks in advancing financial AI
Speaking about risks in advancing financial AI, Qian Bin identified three major risks requiring special attention. First, is there a high-quality dataset and accumulated professional knowledge assets? If a financial institution lacks professional knowledge accumulation based on its own business system as a foundation, model outputs can easily deviate from risk standards, business rules, and operational direction. The key to whether a financial institution uses AI capabilities well lies in whether it can establish high-quality datasets and effectively accumulate professional knowledge based on its own business system. Computing power can be purchased and models can be borrowed, but high-quality datasets cannot be provided by others. Second, whether AI can play a better role depends on establishing end-to-end, systematic, and engineering-oriented implementation capabilities. AI pilot exploration and large-scale application are completely different forms; without unified and complete engineering implementation capabilities, it can easily bring disaster and risk. For example, with intelligent agents, how to control the data, systems, platforms, and related operational permissions they can access, and whether permissions can be effectively controlled and governed across their full lifecycle, is crucial. Third, how to establish a more effective AI governance ecosystem? Any technology is a "double-edged sword" that brings benefits but also risks.
Four areas for strengthening work
Qian Bin believes financial institutions should strengthen related work in four areas. First, set reasonable and clear boundaries for AI innovation by giving it a "steering wheel" and "traffic light" mechanism so that AI explores and experiments within designated lanes. Second, in AI security governance, establish dual-track thinking by both incorporating AI into the scope and objects of security governance and leveraging AI capabilities to improve the level of security governance. Third, the pace of AI evolution and application scale has reached levels difficult for traditional cognition to match, but in high-risk areas, his personal view is that final risk judgments, responsibility bearing, and ultimate value choices should be completed by humans, not simply replaced by machines. Fourth, the financial industry must establish an ecosystem of co-governance. Because the essence of finance is the exchange and circulation of funds, resources, and information, it naturally has contagious and cross-market, cross-institution characteristics; therefore, industry-wide joint defense, regulatory coordination, and global collaboration must be established to build this ecosystem and improve resilience against risks.
He also mentioned that AI applications must not become a new driver of the "digital divide" but should create conditions for financial fairness and equal use of resources, becoming an important force. Finance is originally a scarce resource, so how to make finance more inclusive, convenient, fair, and accessible—AI should play a greater role in this area. Regarding systemic risks brought by AI, Qian Bin said that future systemic risks AI may bring to the financial industry include cybersecurity and data security. In addition, with AI application, market risk and liquidity risk may also arise. When a financial tsunami comes, no single institution can support or face it independently, so global coordination and linkage are certainly needed. He proposed that in the future, effective management of human use of AI must be strengthened. The focus of human work must add certain content: training AI, monitoring AI, handling and intervening in AI anomalies, including final review of AI decisions and bearing ultimate decision-making responsibility.
Comments