China Academy of Information and Communications Technology: AI technology has penetrated into the five major business chains of the financial industry

On January 18, the China Academy of Information and Communications Technology (hereinafter referred to as the "China Academy of Information and Communications Technology") released the "Financial Artificial Intelligence Research Report (2022)" (hereinafter referred to as the "Report"). The report pointed out that at present, artificial intelligence technology has penetrated into the five major business chains of the financial industry such as financial product design, marketing, risk control, customer service and other supporting activities, and artificial intelligence-related technologies such as biometrics, machine learning, computer vision, and knowledge graph have enabled the financial industry, resulting in a number of AI application scenarios in the financial industry such as intelligent identity, intelligent customer service, and intelligent marketing.

According to the statistics of the Banking and Insurance Regulatory Commission, the information technology investment of banking institutions and insurance institutions in 2020 was 207.8 billion yuan and 35.1 billion yuan respectively, an increase of 20% and 27% year-on-year. The banking, insurance, and securities industries are in the wave of digital transformation, and artificial intelligence is one of the technical means of digital transformation, but the application of AI technology should also pay attention to compliance and other issues.

Application scenarios of AI in the financial industry

In 2021, Chinese Min min bank issued the "Specification for the Evaluation of the Financial Application of Artificial Intelligence Algorithms (JR/T 0221-2021)", which stipulates the basic requirements, evaluation methods and judgment criteria for the application of artificial intelligence algorithms in the financial field. Before this specification came out, between 2018 and 2020, there were already many statements in the policy documents issued by local governments to promote the development of financial technology, and there were already many expressions mentioning the application of artificial intelligence technology.

In reality, AI technology has been implemented in the fields of intelligent identification, intelligent claims, intelligent customer service, intelligent risk control, intelligent compliance, intelligent investment research and intelligent investment in the financial industry.

For example, the financial industry is currently one of the main application areas of RAP (Robotic Process Automation). In commercial banks, the automatic verification of treasury tax rebates, automatic filing of unit settlement accounts, automatic reconciliation of clearing funds, anti-money laundering detailed information supplementation, credit card center risk replacement, etc. can be undertaken by RAP; in the insurance industry, RAP can do contract document submission, risk control indicator monitoring and other work; in the securities industry, RAP can be used to achieve monitoring and regular inspection during the opening of the market.

According to the report, Zhongtai Securities has landed RAP solutions in more than ten departments such as retail, credit, and custody, with a total of more than 210 RAP robots deployed, rap running for more than 6,300 hours, and a cumulative conversion cost of more than 4.7 million yuan.

The use of computer vision, speech synthesis, and the collection of real person image, sound, action and other information for model training, virtual digital people can be generated, virtual digital people can be applied in the financial industry reception, guidance business handling and other scenarios.

According to the report, in 2019, Baidu and SPDB jointly created the first virtual digital employee "Xiaopu" in China, and this year SPDB has set up more than ten kinds of digital employee positions, including intelligent customer service, intelligent outbound call, AI marketer, AI lobby manager, etc., creating a labor value of about 2,000 people per year. In 2021, the AI digital employees created by Agricultural Bank of China and SenseTime were officially inaugurated in a business hall of agricultural bank and assumed the role of offline lobby manager.

Intelligent outbound call in the financial industry is more widely used, individual customers are more familiar, the current intelligent outbound call system in the industry has been able to replace the manual completion of multiple rounds of indiscriminate dialogue, the system can use the semantic analysis engine to disassemble customer needs, by matching the words in the knowledge base to reply.

According to the report, Sunshine Insurance introduced voice robots, and from January to September 2021 alone, the service volume reached 1.3568 million, saving 4.8625 million yuan. Chinese Life Property & Casualty Reached more than 1.7 million customers a year through intelligent return visits, of which more than 700,000 people successfully returned visits, and if the efficiency of ordinary human agents dialing 150 phone calls a day for return visits is converted, saving more than 220 people per month in labor costs.

In the insurance industry, smart claims that integrate machine learning, computer vision, intelligent voice, knowledge graphs and other technologies have played a big role. One of the major pain points in the traditional insurance industry's claim settlement process is the low efficiency and high claim cost, and customers are easily dissatisfied.

According to the report, Ping An of China launched the world's first image recognition car insurance flash compensation system for car insurance claims, realizing the loss determination of seconds in taking pictures, and if the amount is automatically underwritten within 2,000 yuan, there is no need to wait for the claims adjuster to arrive at the scene, saving a lot of labor costs and time costs.

For individual customers, voice may be only one of the tools for identification, but for the financial industry, voiceprint may be one of the anti-fraud tools. In the banking industry, voiceprint recognition has been applied to credit card applications in the anti-fraud process, which can block credit card applications by building a voiceprint blacklist.

According to the report, SoundTang Technology helped ICBC build a voice anti-fraud platform, based on the collection, registration, comparison and identification of voiceprints, and provides a basis for risk judgment in credit card applications and due diligence. Marking black voiceprints into storage can deal with fraudulent acts committed by external black intermediaries by continuously packaging them into customers by changing phone numbers and other means.

According to the report, since July 2020, Chinese Life Property & Casualty has identified more than 4,400 cases using the anti-fraud risk identification model, recovering more than 200 million yuan.

Obstacles to the landing of AI technology in the financial industry

Although the financial industry is one of the key markets for AI companies to develop, in the financial industry, AI technology has also found a number of application scenarios, but it does not mean that the landing of AI technology in the financial industry is smooth.

The first is the question of roi. In the early stage of technology introduction, the benefits of AI technology may be difficult to cover the high input costs. Limited by the scale of business, small and medium-sized financial institutions may not have an obvious perception of the improvement brought by AI technology to business processes, and their favor for AI technology is lower than that of head financial institutions.

The second is the problem in the technology development process. If you model one case at a time, the development cycle is long and the reuse rate is low; the scene requirements may change, and the life cycle of the model may be very short without maintenance and iteration. Data is important for modeling, but it may not be shared between different financial institutions.

Again, there is the issue of compliance. Compliance issues are a matter of life and death for financial institutions. As a technology being promoted, artificial intelligence itself has certain risks and hidden dangers, which may not meet the risk control principle of the financial industry. If there is a bias in the training data, the intelligent decisions based on AI algorithms are likely to also be biased, forming data discrimination. The data used in artificial intelligence technology also has the risk of being abused and leaked.

The report advocates that while promoting artificial intelligence to continuously deepen the financial scene and broaden the boundaries of financial business, it is also necessary to carry out diversified governance of trusted artificial intelligence to balance the possible conflict between business transformation and compliance.