There are opinions that finance is one of the best landing scenarios for artificial intelligence (AI). In recent years, financial institutions that have deployed artificial intelligence earlier at home and abroad have tried to apply artificial intelligence throughout the entire business system. For example, the banking industry goes from scenes such as peripheral hall robots, online customer service, and face recognition at the counter, to core processes such as product development, marketing, risk control, and customer service.
Experts believe that although the application of artificial intelligence and other technologies is one of the driving forces to promote the reform of the financial industry, the technology itself, including artificial intelligence, still has certain limitations.
“The improvement of my country’s informatization level has brought the integration of the financial industry and artificial intelligence to an unprecedented height. This is both a challenge and an opportunity for traditional banks. The traditional financial industry has “three highs and one low”, that is, high labor intensity and personnel management. High costs, high business thresholds and low user experience are facing adjustments,” Su Sui, director of 360 Financial Big Data, told the Financial Times reporter.
Su Sui believes that it is the general trend for financial institutions to accelerate the use of big data, cloud computing, artificial intelligence and other financial technology means to reform. This is also the outlet for the financial technology industry to participate in new finance.
“Although people pay more attention to emerging formats and products when it comes to financial technology, technological innovation has long become the main line of financial reform.” said Yang Tao, deputy director of the National Finance and Development Laboratory. “The impact of new technology on finance has penetrated into the financial industry.”
According to reports, since 2016, artificial intelligence was summarized as the three elements of algorithm, computing power, and data. In 2017, the value of the scene was also hotly debated. Regarding this, Su Sui said that AI is only a technology, not a final product, and only when combined with specific businesses and scenarios can its value be brought into play. At present, technologies such as biology and voice have been applied to the financial field on a large scale.
The financial industry has been exploring and practicing how to truly “achieve and implement technology”.
Taking the financial technology industry as an example, 360 Finance has implemented AI practices in intelligent customer acquisition, intelligent marketing, intelligent risk control, and intelligent collection. According to reports, the 360 financial intelligent risk control automated process rate reached 97%, of which the address heat map and the complex relationship network system played a supporting role. Address heat map, which is, relying on the underlying data of the map, through operations such as color marking and upgrading the number of equipment connected in the unit range in the city, combining multiple variables to form GDP information based on each point Comprehensive analysis to determine the size of the customer’s risk. “The darker the color represents the greater the population density. Through business discovery, areas with relatively low population density will have a relatively higher risk.” Su Sui explained.
Yang Tao believes that from the perspective of business scenarios that combine technology and finance, two aspects need to be paid attention to. One is the underlying major technologies, including big data technology, artificial intelligence technology, interconnection technology, distributed technology, security technology, and some still under development. Cutting-edge technology; on the other hand, there are typical financial demand scenarios, such as financial security and financial supervision, payment and settlement, financing products and services, intelligent marketing and service optimization, robo-advisory and wealth management, etc.
Artificial intelligence + the future of the industry
In Susui’s view, artificial intelligence will become an infrastructure like water, electricity, and coal, and companies without AI capabilities will be marginalized.
“From an industry perspective, the future competition is a competition of comprehensive capabilities. Product experience including process and efficiency will become an important measurement standard. The financial industry is showing the trend of artificial, online and intelligent, which will further Solve the problems of the breadth, depth and satisfaction of financial services. From a technical point of view, various behavioral data will be more fully utilized. At present, the value of a large amount of paper-based information accumulated by traditional financial institutions has not been fully explored. The application of structured data will change the structured value of data. In addition, large enterprises and small and medium-sized companies will play different roles in data processing, exploration, and connection.” Su Sui said.
The “Report” believes that in 2019, the artificial intelligence industry is facing a “mid-term exam.” The industry’s requirement for artificial intelligence is no longer “looking up to the stars”, but on a large scale to create new value for society. Regarding the thresholds that hinder the large-scale implementation of the artificial intelligence industry, the “Report” believes that there are four categories: data security and privacy protection, data threshold, talent threshold, and cost threshold.
Yang Tao believes that from a micro perspective, we should not only pay attention to whether the risk characteristics of the original financial institutions and products have changed after the introduction of new technologies, but also discuss the risks of the new technologies themselves and the new financial risks in the Internet and big data environment.
According to industry insiders, artificial intelligence also has certain bottlenecks or loopholes. Yang Zijiang, a professor at Western Michigan University and the founder of Deep Trust, told the Financial Times reporter: “The research on AI has been for a long time, and the reason why there has been no progress for a long time is because previous research is based on semantic understanding. It tries to understand a problem. Respond to the corresponding understanding, but this road is difficult to go down. In recent years, the rapid development of artificial intelligence is mainly based on machine learning. Based on big data, it does not really understand your problem. However, just learning based on big data is not a particularly complete combination. For example, in 2017, Tesla’s self-driving vehicle collided with a truck just one hour after its debut. The reason was that the on-board machine learning software misidentified a white truck as a blue sky. So, it’s not the deficiencies in specific applications, but the deficiencies and bottlenecks in the discipline itself.”
Su Sui said that in theory, there must be a way to crack the loopholes in AI, such as the face and in vivo verification frequently mentioned recently. “But the technology is continuously iteratively improved, and the prevention technology will also be improved. Coupled with hardware upgrades, such as 3D optical lenses, it will be more and more difficult to crack.” Su Sui said.