Summary: From the perspective of the financial sector, the main applications of artificial intelligence in the international banking industry are concentrated in capital operations, market analysis, customer marketing, and risk supervision. Intelligent customer service, smart investment, intelligent quantitative trading... Artificial intelligence has huge room for development in the financial field, but it also makes supervision more complicated.
The rapid development of China's artificial intelligence industry in recent years is welcoming overall planning and comprehensive guidance at the national level. In July 2017, the State Council issued the “New Generation Artificial Intelligence Development Plan†(hereinafter referred to as “Planningâ€). Once the Plan was released, it attracted attention around the world. So, what is the development prospect of artificial intelligence in China? How is it applied in the financial sector? What kind of challenges will the regulatory system bring? This article intends to explore these issues. 
First, the application of artificial intelligence in the financial field
1. Application in the international financial industry
In recent years, the global financial industry has been quietly changing under the catalysis of artificial intelligence. According to the Financial Stability Board (FSB) report, the application of artificial intelligence in the international banking industry is mainly concentrated in the following aspects.
(1) For capital operation, focus on asset allocation, investment research consultants, and quantitative transactions.
The use of artificial intelligence in financial investment advisors, often referred to as smart investment, mainly refers to providing customers with algorithm-based online investment advisory and asset management services. Specifically, it can be divided into three categories: First, it is applied to the sales front-end of large-scale asset allocation-type smart investment, mainly through user analysis to solve large-scale asset allocation problems for customers, such as Wealthfront; second, investment in investment analysis stage Research-based intelligent investment, mainly through massive data mining and logical chain to solve investment research problems, such as Kensho; Third, intelligent quantitative trading system applied to strategy, transaction and analysis, mainly through artificial intelligence to replace traders, applied to investment Trading, such as WaterBridge's 24/7 artificial intelligence trading. According to the statistics company Statista's forecast, in 2017, the US smart investment management assets will reach 224.802 billion US dollars, and will reach 509.55 billion US dollars by 2021, with a compound annual growth rate of 29.3%.
(2) Market-oriented analysis, focusing on trend forecasting, risk monitoring, and stress testing.
Artificial intelligence technology can obtain more information from scattered historical data, help identify nonlinear relationships, give market forecasts (price fluctuations) and their timeliness, resulting in higher returns, directly or indirectly. In addition, artificial intelligence technology can analyze large, semi-structured and unstructured data sets, taking into account changes in market behavior, regulatory rules and other trends, conducting reverse testing, model validation and stress testing to avoid underestimating risk To increase transparency. For example, the world's first purely artificial intelligence-driven fund, Rebellion, predicted a stock market crash in 2008 and rated the Greek bond F in September 2009, one month ahead of Fitch. Senoguchi, a machine invented by Mitsubishi Corporation of Japan, predicts that the Japanese stock market will rise or fall after 30 days on the 10th. After 4 years of testing, the correct rate of the model is as high as 68%.
(3) Customer-oriented marketing, focusing on identification, credit evaluation and virtual assistants.
Artificial intelligence technology has been widely used in the financial front, and large customer data has been imported into the chat program, enabling it to communicate face-to-face with natural language and improve the ability to “get customersâ€. In April 2017, Wells Fargo began piloting a chat bot project based on the Facebook Messenger platform. The virtual assistant communicated with users to provide account information to customers and help customers reset their passwords. And Erica, the intelligent virtual assistant of Bank of America, also officially unveiled. Users can use voice and text to interact with Erica. Erica can help users check credit scores, view consumption habits, and provide repayment advice and financial guidance to more than 45 million customers as the bank's revenue and expenditure changes. In addition, HSBC has used face recognition and voice-based biometrics to verify consumer identity; Royal Bank of Scotland uses the “LUVO†virtual dialogue robot to obtain the most suitable home loan for customers, etc., designed to become “trustworthy†for users. Financial consultants."
(4) For financial supervision, focus on identifying abnormal transactions and risk subjects.
Artificial intelligence technology can be used to identify anomalous transactions and risk entities, detect and predict market volatility, liquidity risk, financial stress, housing prices, industrial production, GDP, and unemployment, and capture threats that may pose a financial stability. Currently, some international regulators, such as the Australian Securities and Investments Commission (ASIC), the Singapore Monetary Authority (MAS), and the US Securities and Exchange Commission (SEC), use artificial intelligence for suspicious transaction identification. Specific practices include identifying and extracting stakeholders from evidence documents, analyzing users' trading trajectories, behavioral characteristics and associated information, and cracking down on criminal activities such as underground money laundering more quickly and accurately.
2. Application in China's financial industry
In China, the banking industry has also followed the pace of the international banking industry and began to explore the application of artificial intelligence technology, in which Internet finance companies have seized the leading edge in the research and application of artificial intelligence. For example, Ali's Ant Financial has applied artificial intelligence to Internet small loans, insurance, credit reporting, asset allocation and customer service and achieved good results. Tencent U-Map is a face detection application of Tencent. It also cooperates with Tencent Credit, Weizhong Bank and Tenpay to realize the credit evaluation of users.
(1) Intelligent customer service. At the end of 2015, Bank of Communications launched China's first intelligent artificial intelligence service robot “Jiaojiaoâ€, which has been employed in business outlets in Shanghai, Jiangsu, Guangdong and Chongqing. This robot adopts the world's leading intelligent interactive technology, and the interactive accuracy rate is over 95%. It is the first intelligent service robot in China that can truly listen and speak and judge. On the basis of the “Enterprise Access†platform, ICBC uses data docking and smart devices to optimize business processes and innovatively launch self-service account opening services for public customers. Customers only need to go to the outlet once to complete account opening and settlement of products. , data printing, reservation printing and other business processing.
(2) Smart investment. At present, there are many companies providing this service in China, including banking departments (such as Guangfa Zhitou, China Merchants Bank Motech Zhitou, ICBC “AI†investment, etc.), and fund departments (such as Southern Fund Super Zhibao, GF Fund Jizhi Finance, etc.). Large Internet companies (such as Baidu Finance, Jingdong Zhitou, Tonghuashun and third-party startups (such as Mi Cai, Blue Ocean Fortune, Latte Finance, etc.) have applied in smart investment.
(3) Intelligent quantitative trading. Under the current financial supervision system in China, banks have relatively few applications in this area, but Jingdong Finance, Ant Financial Services, Keda Xunfei, and causal trees have actively explored. For example, causal trees use machines to automatically select high-quality projects and launch supernovas every week, helping companies to increase their probability of getting the next round of financing in the next six months to around 30%. The Harvest Fund has developed a comprehensive investment decision system, the Harvest FAS System, from market forecasting, asset allocation to product selection, and achieved a return on investment that exceeds the market's yield.
(4) Risk control and management. This mainly includes the following three aspects: one is data collection and processing; the second is risk control and forecasting model; the third is credit rating and risk pricing. For example, a traditional loan business may take two to three days to approve, and an automated approval model based on an artificial intelligence model may take only a few seconds to complete, while some traditional wind control models may take several months to iterate. Even years, but artificial model iterations can be very convenient and automatic. Bank of China launched the anti-money laundering verification project for trade finance business, using artificial intelligence technology such as text analysis, image recognition and machine learning to reduce the original audit time from 2 hours to 2 minutes. The efficiency and quality have been greatly improved. Labor costs are greatly reduced.
3. The development space of artificial intelligence in the financial field
(1) Enhance the ability of financial institutions to attract customers and gain market competition initiative.
The rapid development of artificial intelligence enables the machine to simulate human functions to a large extent, to achieve mass and personalized service to customers, which will have a profound impact on the financial industry at the high end of the service value chain. Artificial intelligence will It becomes an important means for banks to communicate with customers and discover their financial needs, thereby enhancing the bank's stickiness to customers. It will bring a new round of changes to financial products, service channels, service methods, risk management, credit financing, and investment decisions. Artificial intelligence technology can be used to serve customers at the front end, support decision-making in credit, various financial transactions and financial analysis in China, and be used for risk prevention and supervision in the background. It will greatly change the existing financial structure and financial services. Personalized and intelligent.
(2) Reduce the operating costs of financial institutions and improve work efficiency.
Financial institutions can use artificial intelligence and machine learning to develop new business needs, reduce costs and manage revenue risks, improve operational efficiency, and optimize customer processes. According to the "2016 China Banking Service Improvement Report" issued by the China Banking Association, in 2016, banking financial institutions lost 177.714 billion transactions, a year-on-year increase of 63.68%; the banking industry's exit rate was 84.31%, year-on-year. Increase by 6.55 percentage points; the transaction amount from the counter is 1522.54 trillion yuan. Among them, 15 banks have exceeded 90% of the business. In the future, more and more financial institutions will join the process of using artificial intelligence to enhance their competitiveness.
Authorities and experts are generally optimistic about the application of artificial intelligence in the financial sector. BenGoertzel, president of the Artificial Intelligence Society, believes that artificial intelligence may be involved in most of the world's financial transactions after 10 years. ATKearney, an overseas consulting firm, expects robotic investment advisors to become mainstream in the next three to five years, with a compound annual growth rate of 68%. By 2020, the asset size managed by the Robotics Investment Advisor is expected to reach $2.2 trillion. Citibank Research predicts that assets managed by artificial intelligence investment consultants will achieve exponential growth over the next 10 years, totaling $5 trillion. Deloitte pointed out in the report "Banking Outlook: Banking Remodeling" that the application of machine intelligence decision-making will accelerate development. Intelligent algorithms will become stronger and stronger in the process of predicting market and human behavior. Artificial intelligence will affect industry competition and the market will become more efficient.
Second, the supervision of artificial intelligence in the financial field
1. Regulatory objects tend to be complicated
In the current regulatory system, the targets of supervision are often legal persons and natural persons. Due to the development of artificial intelligence technology, the owners and operators of investment accounts may change. For accounts with ownership as the main body of the collection, the principle of penetration will be difficult to trace back to the subject. This is because the actual controller is not a subject but a smart agent. Thus, the challenge of regulation is complex. Investors believe that the account is not operated by any of them, and the actual controller is not theirs. The intelligent agent service provider only provides the intelligent agent "product" and does not actually control the account. At this time, the regulatory authorities have to face the problem of how to supervise the "smart agent" that is neither a natural person nor a legal person.
2. It is difficult to identify violations of laws and regulations
For example, a large number of investors hire the same high-performing smart agent to manage their own account investments. Because the operation logic of the same intelligent agent is similar, these accounts are legally independent and not related, but their actual operation may be characterized as “coherent action peopleâ€. Therefore, even if the big data analysis system of the regulator can “capture†this phenomenon very sensitively, how to determine this “same hero†behavior will be a regulatory problem.
3, intelligent agent behavior increases the difficulty of supervision
Although technically speaking, intelligent agent behavior can be controlled from the internal control procedures, the current regulatory regulations are not involved in the regulatory boundaries and responsible subjects of their specific agency behavior.
4. The subject of responsibility is difficult to define
If individual developers design a malicious intelligent agent and are used by some pooled funds, it may trigger a change in the price of individual stocks. For such violations, existing regulations will be difficult to define the subject of responsibility.
Third, the development of artificial intelligence in the financial field
Accelerating the application of artificial intelligence in the financial field is the future development direction. Regulators must face this trend and actively seize the high ground for artificial intelligence development. They must also pay attention to the impact of artificial intelligence application on the financial sector, and carry out forward-looking research and strategy in a proactive manner. Sexual deployment.
1. Research and improve financial market trading rules for the characteristics of artificial intelligence
China's market trading rules for the application of artificial intelligence finance are almost blank. We should actively study the trading rules of relevant financial markets in order to create a good market environment for the development of artificial intelligence.
2. Accelerate the application of artificial intelligence in financial supervision
The application of artificial intelligence in the financial field has put forward new requirements for financial supervision modes and means. Faced with the rapid development of artificial intelligence, China's financial regulatory authorities should actively introduce artificial intelligence to further improve regulatory efficiency.
3. Pay attention to the protection of user privacy
At present, the legal system of privacy protection in China is still not perfect. The awareness of privacy protection of financial consumers is relatively weak, and the phenomenon of personal information leakage has occurred from time to time. Whether it is to protect the basic rights of citizens or the need to develop artificial intelligence, It is necessary to improve China's financial privacy protection system, strengthen relevant administrative supervision, clarify the financial institution's relevant notification obligations, information security guarantee obligations, and liability after problems occur, and effectively ensure the information security of artificial intelligence in the financial field.
For commercial banks, first, large financial groups should do a good job in the early stage of capital technology, intervene in advance, and strengthen technological innovation; accelerate business innovation, maintain a leading position in industry transformation, and strengthen technology and maintenance personnel reserves, especially The introduction and cultivation of intelligent and compound talents will enhance their core competitiveness and meet the development requirements. The second is to strengthen risk control. In terms of data processing, artificial intelligence technology has greatly expanded the data source, so more data is included in the analysis system. At the same time, financial instruments can automatically evolve trading strategies and even simulate experts to make decisions, which implies many new risks. Strict review of the previous data sources and intelligent program design must be carried out to strengthen risk control. In particular, in the extreme cases of terrorist attacks, regulatory changes and the implementation of short-selling bans, experts are also required to carry out the necessary risk detection and response.
C&Q Technology (Guangzhou) Co.,Ltd. , https://www.gzcqteq.com