At the 22nd China International Finance Forum held in Shanghai on December 19-20, Li Lihui, former President of Bank of China, delivered a keynote speech on building a smart financial ecosystem in the era of digital economy.
Li emphasized that the potential security risks and technical flaws of artificial intelligence (AI) have not been mitigated by the algorithmic innovations of generative AI. Given that finance is an industry with stringent demands for security and trustworthiness, it is imperative to safeguard financial assets and data, ensure the reliability of transactions and services, and maintain the accuracy of accounting processes and records.
Based on current practices, Li outlined three key requirements for smart financial innovation in the near to medium term: high reliability, explainability, and cost-effectiveness. "The cornerstone of smart financial innovation is trustworthiness," he said. "We must balance security and efficiency, ensuring that algorithms align with real-world scenarios to build credible models that earn the trust of clients, markets, and regulators."
Li stressed that smart financial innovation is not merely about overlaying traditional systems with AI but involves fundamentally reforming institutions, redesigning processes, and rebuilding underlying systems. He advocated for a balanced regulatory approach in smart financial governance—one that is neither too rigid nor too lax—to foster innovation while mitigating risks.
**Key Insights from the Speech:**
1. **Financial Models: Reliability and Cost-Effectiveness** AI advancements, particularly multimodal models, enable financial institutions to process unstructured data (e.g., text, audio, video) and enhance customer interactions. However, challenges such as model hallucinations, biases, and privacy leaks persist. Financial models must prioritize security, transparency, and scalability to reduce development costs and broaden applications.
2. **Financial Agents: AI Substitution and Legal Status** AI agents are evolving from assistants to autonomous decision-makers, capable of handling high-value tasks like market analysis, risk assessment, and investment advisory. Their deployment in banks, insurers, and asset managers is reshaping workforce structures, displacing both labor- and knowledge-intensive roles. Legal frameworks must define their boundaries, responsibilities, and accountability.
3. **Data Sharing: Quantity and Quality** As a data-intensive sector, finance relies on robust data ecosystems. Current gaps include fragmented public data, inefficient private data flows, and underutilized behavioral datasets. Li called for open public data platforms, anonymized private data sharing, and specialized financial databases to unify and elevate data standards.
4. **AI Competition: Hardware and Software Capabilities** The global AI race, led by China and the U.S., hinges on computing power ("hard" and "soft" capabilities). While the U.S. prioritizes hardware, China’s dual-track strategy—exemplified by breakthroughs like DeepSeek-V3’s resource-efficient algorithms—showcases its potential to rival global leaders. However, geopolitical tensions and technological sovereignty remain critical challenges.
Li concluded by underscoring the importance of balancing intellectual property rights with open-source collaboration to democratize AI access while safeguarding national interests. He expressed confidence in China’s ability to cultivate a secure, efficient, and innovative smart financial ecosystem.
Comments