From bancassurance business to securities market, AI empowers the financial industry

On July 7, CBN invited Yang Shoujun, chief strategist of Shanghai Huizheng Finance, Gu Yuanzhong, researcher of Huizheng Institute of Finance and Economics, and Ling Xufeng, dean of the School of Artificial Intelligence of Tianhua College of Shanghai Normal University, to discuss how artificial intelligence affects the financial industry. During the discussion, the three guests discussed "AI+" on traditional business and investment business in the financial industry.

Empower traditional bancassurance business

Regarding the impact of artificial intelligence technology on the financial industry, Gu Yuanzhong believes that the big prediction model represented by ChatGPT launched by Open AI marks a big step forward in generalized artificial intelligence. "In fact, the empowerment of artificial intelligence for our financial industry has actually begun a long time ago, and it is a silent process in the bit by bit. For example, AI is helping banks' telephone customer service become smarter, from menu-based to inquiry-based. Gu Yuanzhong said.

Gu Yuanzhong also believes that artificial intelligence will be further applied to the intelligent customer service of banks in the future. For example, intelligent customer service can identify the meaning of different languages in the same context, from "I want to check the balance" to "how much money is left" and other questions, artificial intelligence can accurately identify keywords. Even, intelligent customer service will actively push the bill according to the user's habits. He said that such financial services changes are based on ChatGPT, which breaks the traditional operating model of conversational bots to exhaust this user's conversational intent. From a more natural language perspective, to generate this extremely rich range of dialogue content and responses. "In the past, he may have done multiple-choice and judgment questions, and he may be doing essay questions in the future."

In Yang Shoujun's view, artificial intelligence is more empowering for the insurance industry, first, it is easy to solve claims, and after the claims work is simple, it can play different claims results for different insurance customers and improve the service effect; The second is to improve the efficiency of the use of freight insurance, shape the portrait of customers for insurance companies, and adopt corresponding pricing levels and rates for different types of customers' return behavior.

"For example, a customer is a brand user, never returns after buying, freight insurance can be sent for free, maybe you buy 100 times and refund 1 time, the cost is completely borne by the insurance company. If a customer buys ten items and wants to return eight items, and often returns them, the specific pricing and rate may have to be considered differently. Yang Shoujun explained.

Ling Xufeng pays more attention to the application of artificial intelligence in financial product design. He believes that artificial intelligence does product design with three characteristics. First, refined design based on big data that grasps user characteristics and preferences can be carried out. Second, generate personalized design solutions according to the potential needs of customers. Third, continue to use the power of machine learning to improve product design.

New opportunities in the securities market

Regarding how AI can further change us in the financial industry, the three panelists turned their attention to the securities market. For Ling Xufeng, according to the large model deep learning ability, abstraction ability, reasoning ability, and zero-sample learning ability, it can play a subversive role in many positions in the financial field. "For example, the valuation of stocks and bonds, intelligent customer service, intelligent investment advisory, programmatic trading, risk control, etc., and even a very important one, I think in terms of investor education or financial practitioner education." Ling Xufeng said.

Specific to intelligent customer service and intelligent investment advisory, Ling Xufeng compared two large models in the financial field, one is the closed-source, fee-based BloombergGPT, and the other is the free, open-source FinGPT. He said that both have two advantages, one is 24 hours a day, and the other is that they can handle large amounts of financial data. In addition, he pointed out that AI is less affected by subjective bias, which is an advantage that humans do not have, and in terms of intelligent customer service and intelligent investment advisory, it does not have the pressure of marketing business and can give neutral guidance to customers.

Whether for institutional or individual investors, Gu Yuanzhong believes that AI has advantages in predicting stock prices, first of all, AI analyzes a wider range of data, including fundamental factors, technical factors, and unstructured data, and it will mine some trends that humans ignore; Second, AI's automated models can improve the speed of prediction; Finally, AI can lower the threshold for stock research and save the cost of research.

As a researcher, Gu Yuanzhong pointed out that AI can really liberate productivity for researchers: First, AI can carry out the most basic document collation, such as the minutes of industry seminars, which can be directly recorded into text, or further extracted views; Second, AI can track large-scale research reports, capture basic data, and generate automated reviews; Third, AI can help researchers with knowledge management and help us build our own knowledge warehouse.

At present, the securities market has entered the era of comprehensive registration system, and Yang Shoujun believes that AI can avoid risks for us in the market. He said that it can be considered from two dimensions: first, in the early stage of the listing of a listed company, it can help investors discover its potential risks; Second, for some listed companies that have been poorly operated, AI can help investors screen companies with financial risks.

"I just talked about the relationship between Chinese medicine and Western medicine, you look at it every day, 5,000 stocks you look at and hear, and you can't finish looking at it all day. But with artificial intelligence, it may also be an empowerment for our work, and for me, we can give investors more professional advice, and through quantification, some of the original measures are simplified. He spoke bluntly about the advantages of using AI technology.

Yang Shoujun also focuses on the special applications of AI. He proposed that AI could help individual investors draw their own "investment electrocardiograms." Each individual investor has his own trading trajectory, but each operation may not have been analyzed, specific to the buying and selling point, what the mood was at that time. With a customized AI model, it can serve as a reminder to investors during the trading process, and the development of the model should be visible in the short term.

Huang Wei, the host of First Finance and Economics, said that in the foreseeable future, AI is bound to fully empower investor management, and reduce market volatility by weakening information asymmetry, and so on, and ultimately promote the entire stock market to effectively improve. "Of course, the application of AI is a double-edged sword, and while AI brings huge development opportunities to the securities market, it may also produce technical risks, legal compliance risks and other issues. Therefore, in fact, we must carefully look at the two sides, opportunities and risks that AI can bring, and control the steering wheel of AI application in investment transactions to ensure that transactions eventually and reliably drive into the future. Huang Wei finally concluded.