The AI Wave Sweeps Over the Banking Industry: Opportunities and Challenges Coexist

The AI Wave Sweeps Over the Banking Industry: Opportunities and Challenges CoexistBy Zhang Hong and Li Yuwen, Editors: Ma Ziqing"The number of banking jobs that AI might replace could be greater than in any other sector." According to a recent report by Citigroup, about 54% of jobs in the US banking industry are potentially automatable, and a further 12% can be enhanced through AI

The AI Wave Sweeps Over the Banking Industry: Opportunities and Challenges Coexist

By Zhang Hong and Li Yuwen, Editors: Ma Ziqing

"The number of banking jobs that AI might replace could be greater than in any other sector." According to a recent report by Citigroup, about 54% of jobs in the US banking industry are potentially automatable, and a further 12% can be enhanced through AI.

This prediction is not baseless, as current trends in the banking sector reflect this view. Especially in the past two years, with the emergence of generative AI large models, AI+Finance has developed rapidly. In terms of application scenarios, banking institutions have explored and matured AI applications in various areas, including customer service and intelligent Q&A, credit approval and risk management, intelligent operations, and process automation.

Financial Institutions Embrace AI: A Necessary Path of Self-Transformation

In the current environment of persistent interest rate pressure and slowing revenue growth, financial institutions, with the banking industry as a representative, are undergoing a process of self-transformation "cutting to the bone." AI innovation in finance is driving financial services towards greater intelligence and personalization, enabling financial institutions to manage risks more effectively, enhance operational efficiency, and optimize customer experience.

The Status Quo of AI Application in the Banking Industry: Blossoming Everywhere

"In the banking industry, all commercial banks are embracing AI technology, hoping to leverage technological innovation to support business innovation and development, enhance service levels, and stand out in differentiated competition," said a relevant official from Nanjing Bank.

Due to the sheer volume of customers and the maturity of AI technology, retail AI scenarios are more common in commercial banks, such as intelligent Q&A, OCR-assisted data entry, anti-fraud, and intelligent investment advisors.

Various AI application scenarios are also gradually emerging among small and micro enterprises, medium and large enterprises, groups, and financial market clients. AI scenarios that enhance efficiency and quality are also gradually appearing in the internal management links of banks.

Applications of AI in Banks Financial Market Operations: Quantitative Trading Algorithms and Virtual Traders

An employee working in risk management at a state-owned bank told reporters that AI is primarily used in quantitative trading algorithms and virtual traders in banks financial market operations.

For example, in client-agent trading of precious metals and foreign exchange, operations are repetitive and structured, often differing only in direction and amount. Therefore, quantitative trading algorithms (AI) can be used to replace traders in executing market strategies.

In fund operations, virtual traders can leverage generative AI to communicate, quote, and even complete transactions through chat in over-the-counter transactions such as fund lending.

AI Simplifies Loan Approval Processes: Enabling Fast Approval of Easy-Loan Products

Banks are also simplifying loan approval processes by applying AI technology to easy-loan products. Using AI to analyze customer information creates a whitelist where customers on the list will get credit limits quickly after submitting requests.

"Traditional approval processes are cumbersome," said the employee from the state-owned bank. "They require credit officers to communicate with customers, sign materials, and submit applications to branches or headquarters, which can take weeks or even months. 'Easy-loan' products make approval process model-based and real-time, enabling customers to receive feedback promptly."

AI Applications in Settlement Operations: Anti-Fraud and Anti-Money Laundering

AI is also used for anti-fraud and anti-money laundering in settlement operations, predicting transaction risks. "Previously, an expired ID card required renewal at a branch," noted the state-owned bank employee. "Now, taking a photo for facial recognition is enough without visiting a branch, thanks to AI applications."

Generative AI Empowering Financial Services: Providing Personalized Services and Reducing the Risk of Privacy Information Leakage

Li Mingshun said that generative AI is being used to provide lightly guided services, including investment and wealth management consulting for high-net-worth clients. Combined with digital human technology, this guides customers through procedures, provides professional investment advice, and reduces privacy information leakage risks.

In traditional human services, customer managers may remember sensitive information, posing a risk of information leakage. AI effectively protects customer privacy and prevents data leaks.

AI-Assisted Collection: Improving Collection Efficiency and Avoiding Improper Communication

AI can also be applied to debt collection. Li Mingshun stated, "In the past, debt collection was typically performed manually. However, manual collection involves emotional fluctuations, which could lead to improper communication and extreme behavior. Generative AI-assisted debt collection can avoid extreme language through preset expressions, ensuring politeness and professionalism without the need for rest.

Silicon Valley Venture Capital Firms Play with AI: From Project Screening to Investment Advice Writing

Participating in project screening, due diligence... Silicon Valley venture capital firms play with AI.

Discussing future explorable areas, the state-owned bank employee mentioned that AI can also be used for post-loan management, warning of market risks, credit risks, or other risk indicators. Additionally, due to the inherent advantages of generative AI in text processing, future applications in public opinion management can be explored.

Li Mingshun said that AI-assisted interviews hold great potential in large-scale recruitment. AI can accurately record a candidates emotions, personality, and professional level, helping interviewers objectively score and reduce the influence of personal relationships. Some companies have already tried AI interviews in chain enterprises, although not directly serving finance, but with similar principles, indicating broad prospects for AI applications in financial human resources.

Advantages of AI in Financial Services: Increased Efficiency and Reduced Costs

Numerous interviewees pointed out that the advantages brought by AI in finance are multifaceted, with the most prominent being efficiency and cost optimization. Compared to traditional business operation processes, AI technology can automate and intelligently process massive amounts of data, enabling fast and accurate decision-making, thereby significantly saving time and manpower costs.

AI Enhances the Precision of Financial Applications: Promoting the Popularity of Financial Services

Tinus, head of innovation and development at Tongdun Technology, told reporters that AI's involvement can improve the accuracy of applications like financial risk control and financial marketing, boosting financial institutions efficiency, reducing operational costs, and enhancing the popularity of financial services.

AI Empowers Financial Institutions Digital Transformation: Enabling More Efficient Decision Support

Tinus further explained that the integration of AI with big data, cloud computing, and other technologies expands the application of massive data and supports more complex and intelligent financial intelligent applications through powerful computing capabilities. The fusion of AI and blockchain technology leverages blockchains distributed ledger, smart contracts, and other capabilities to achieve data immutability, traceability, and enhanced trust in financial data applications, increasing the reliability and manageability of AI applications in finance. Moreover, the integration of AI and cryptography technology breaks down the security barriers for data intelligent collaboration and sharing across institutions, expanding the boundaries of AI data applications, particularly addressing the challenges of private data sharing and circulation in the financial field. Technology solves the problems of secure, reliable, and intelligent circulation of financial data.

Trustworthy AI: Building the Future Foundation for Financial Technology Security

Tinus believes that data security throughout its lifecycle is crucial for AI applications in finance, and trustworthy AI will become a significant development trend. Trustworthy AI features include security, robustness, fairness, explainability, privacy, and controllability. Combining and integrating technologies like privacy computing, blockchain, and data sandboxes are effective ways to achieve privacy-enhanced AI. Building a secure foundation for trustworthy AI through privacy enhancement will empower financial technology to transition from "digitalization" to intelligent digitization, enabling secure, reliable, and traceable operations to promote sustainable development and further innovative applications of AI.

AI Agent: A New Direction for Future Financial Technology Applications

"AI Agent is a significant trend in AI applications in the financial field based on large models. AI Agent is built upon large language models while possessing features like memory retrieval, decision reasoning, and action sequence selection, making it more advantageous in handling complex tasks and considered a future direction. AI Agents intelligent decision-making, automated execution, personalized service, continuous learning, and optimization characteristics can provide financial institutions with more intelligent and efficient services and decision support," said Tinus.

Challenges of AI Application: Compliance Risks, Privacy Protection, and Market Vulnerability

Despite the vast potential of AI applications, several pain points exist in its current use in finance:

• Potential for Market Vulnerability: If financial market participants widely adopt the same model for decision-making, it could lead to increased market vulnerability. If all participants use the same model and input the same data, they are likely to reach similar conclusions. Financial market transactions require agreement between buyers and sellers with differing opinions. If most participants hold the same viewpoint, such as simultaneously choosing to sell, market fluctuations could become significantly more volatile.

• Large Models Not Yet Suitable for Core Departments in the Financial Industry: Li Mingshun summarizes the difficulties of applying large models in the financial industry. He notes that applications of large models in finance can be understood as a branch of AI technology. Large models primarily find use in marketing and service, such as data organization and consultation services in the financial sector. However, their application is not straightforward in more stringent and critical financial domains such as risk control and asset pricing because large models currently suffer from hallucinations. Li Mingshun believes, "Large models are essentially language models capable of understanding and processing text, but not rule engines. They are not suitable for the core departments of the financial industry.

• Lack of Standardized Institutions in the Personal Credit Scoring Field: Li