Not long ago, the global anti-money laundering standards-setting bodies - the Financial Action Task Force (FATF) published the "Chinese anti-money laundering and anti-terrorist financing mutual evaluation report" positive progress made by China in recent years, anti-money laundering work recognized that China's anti-money laundering system have a good foundation. Among them, the role of artificial intelligence has become increasingly prominent
Money laundering refers to the proceeds of crime resulting drug, organized crime syndicates, and terrorist criminal activities, smuggling, crime, corruption and bribery undermine the financial management order crime, financial fraud and produced by a variety of means to cover up financial institutions, conceal the source and nature of the funds, it is formally legalize behavior. At present, the common money laundering involving a variety of ways to a wide range of banking, insurance, securities, real estate and so on.
In order for anti-money laundering work more rigorous and detailed, in October 2018, People's Bank of China, Bank Insurance Regulatory Commission, Securities Regulatory Commission jointly issued the "Internet practitioners financial institutions anti-money laundering and anti-terrorist financing (Trial)"; February 21, 2019 silver CIRC issued Decree 2019 of the first "banking financial institutions anti-money laundering and anti-terrorist financing management approach"; and so on.
All along, to carry out anti-money laundering banks, mainly rely on anti-money laundering expert rules of thumb. But with banks trading volume increased year by year, the number of cases of suspicious transactions grew 30-40%, to reduce the amount of suspicious cases only by artificial rules optimization, optimization is difficult to establish the rule of long-term mechanism.
It is noteworthy that, researchers have used artificial intelligence machine-learning platform developed anti-money laundering suspicious transactions intelligence analysis platform that provides decision support anti-money laundering is driven by artificial intelligence machine learning algorithms for anti-money laundering authorities.
The new money laundering trends and challenges
Economic Daily News reporter Le Jiedao, Money Laundering activities there are some new trend, which has brought new challenges for the anti-money laundering. Many in the industry view, the face of the grim situation of anti-money laundering, anti-money laundering supervision need to increase.
At present, the case of the People's Bank from money-laundering cases uncovered point of view, one of the trends of money laundering activity is spreading to the developed regions. Specifically, money laundering predicate offenses (drugs, triads, smuggling gangs, etc.) gradually spread to the inland, largely due to the inland areas relatively relaxed environment facilitates money laundering. The financial regulation in economically developed regions continue to strengthen, money laundering criminals living space squeezed.
At the same time, increase stakeholder cases of crimes committed. According to reports, money-laundering predicate offenses covered by the economic crimes, drug crimes, smuggling and crime in crime extend to crimes committed stakeholder type cases. Crimes stakeholders Illegal pyramid schemes, illegal fund-raising as well as a variety of "black funds" and other cases abound. The crimes committed mostly concentrated in bank credit, securities and futures, transportation and other fields.
In addition, of particular note it is that, as the financial and non-financial business transactions constantly updated, money laundering is becoming increasingly complicated and confusing. When online banking, electronic money, electronic trading occurs, money laundering proficient in electronics technology trail became more erratic. While the development of Internet technology, money-laundering techniques escalating. Internet lending platform, Internet safety, virtual currency, money laundering have become a breeding ground for criminals. Currently, the use of electronic payments, electronic trading network into the mainstream of money laundering, money laundering occur on a large number of online. Data black production and money laundering gang association, application fraud, transaction fraud, money laundering boundaries become increasingly blurred.
"The development and implementation of anti-money laundering work, is of great significance for the orderly development of China's economic and social health: not only conducive to the timely tracing and confiscation of proceeds of crime, to curb money laundering and predicate offenses, maintain economic security and social stability, and facilitate eliminate the potential financial risks and legal risks of money laundering to financial institutions brought maintain financial security; at the same time, be able to cut off funding sources and channels of funding of criminal behavior, prevention of criminal activities, protection of property rights of the victims, to maintain the dignity of law and social justice. "Zhongguancun president of Internet financial Research Institute, Yong said, actively participate in international cooperation in anti-money laundering, but also to maintain good international image of our country.
AI to improve monitoring capabilities suspicious cases
In terms of prevention, monitoring money laundering, banks and other financial institutions based on customer identification, large transactions, suspicious transaction reporting and record-keeping system as the core content, etc., to achieve the goal by funding anti-money laundering monitoring. Way from the point of view, to carry out anti-money laundering banks, mainly depends on the experience of experts in anti-money laundering rules.
However, relying solely on anti-money laundering workflow rule of thumb, the problems encountered in the current anti-money laundering regulatory environment has become increasingly evident. A bank employees said the biggest difficulty is the rapid increase in trading volume led to cases doubled the contradiction between the limited human resources, how to more accurately and efficiently identify suspicious money laundering transactions, banks need to be resolved The problem.
For example, each year a significant increase in suspicious transactions to reduce the amount of suspicious cases only by artificial rules optimization, optimization is difficult to establish long-term mechanism of rule. The identification of suspicious transactions only randomly assigned to investigators, the investigation of cases can not be a rational allocation of tasks based on seniority investigators working with the best time.
In another example, the description of suspicious cases depend on artificial summary, which affect the reported cases of process efficiency and information of the case review process manageability, traceability. In this regard, large financial institutions need to improve management efficiency review of anti-money laundering. Large financial institutions trading large base, a large number of suspicious transactions alarm system, and after the manual review report rate, cost a lot of manpower audit costs, an urgent need for a "risk-based" principle, guide and optimize the allocation of resources against money-laundering.
"Large banks are generally equipped with several hundred anti-money laundering manual review team. The use of AI (artificial intelligence) anti-money laundering technology can help financial institutions while controlling risk, saving more than 30% of the audit workload, equivalent to an annual saving of tens of millions of human cost. "fourth Paradigm technology Co., Ltd., vice president of Chai Yifei representation, AI can also improve the anti-money laundering monitoring capabilities suspicious cases, false negative cases supplement to help customers avoid unnecessary regulatory penalties, and legal, reputation and operational risk.
Anti-money laundering rules iterative optimization system
"Artificial intelligence technology is a systematic project can be iterative, will be fully integrated supervised and unsupervised algorithms to detect new features of money laundering, money laundering and suspicious cases were iterative optimization evaluation model; In addition, the combination algorithm to detect the newly acquired feature , the use of multi-classification model, but also intelligent recognition type anti-money laundering and to strengthen the anti-money laundering suspicious evaluation model, to achieve major identify suspicious cases. "Chai Yifei said the artificial intelligence model can nurture anti-money laundering knowledge to achieve the accumulation of knowledge, the anti-iterative optimization money laundering system of rules, and implement anti-money laundering audit loop optimization system.
So, how to build anti-money laundering suspicious transactions intelligence analysis platform? Reporter in an interview that the specific can be divided into four stages.
First, identify the precise stage, which is established by reviewing the case file AI recognition model, with high dimensional feature a more precise characterization capabilities, significantly enhance anti-money laundering recognition accuracy rate, pinpoint suspicious case file, optimizing resource allocation reviewers.
Second auxiliary audit stage, i.e. by the model feature recognition model of artificial intelligence technology, business and engineering through parsing, transforming into a supporting information for supporting determination of the case file. At the same time, data modeling through data processing, but also become an important trading case file statistical analysis of information.
Phase III is the precipitation of knowledge, namely the introduction of more artificial intelligence, AI assistance to strengthen the capacity of the business. For example, the introduction of technical knowledge map, by third party beneficiary identification data, extract information data collation, targeted for effective use. The introduction of NLP techniques and character recognition technologies to provide early warning needs analysis monitoring of money laundering.
Phase IV is to explore new knowledge, this is the AI model iterative exploratory stage, the establishment of a multi-classification model in a transverse dimension to focus on the kind of the kind of anti-money laundering, classified by AI models. In the vertical dimension, the refinement of anti-money laundering suspicious cases recognition model, through the accumulation of enough data to build customer money laundering risk classification model.
It is worth mentioning that, industry experts said that despite the huge potential of AI technology, but is still one adjunct anti-money laundering, in practice, the financial services industry also needs to better understand the risks and limitations of artificial intelligence .
Dalian abortions cost mobile.120dlbh.com