Fubon Financial

The First AI Anti-Money Laundering Model in Taiwan & Setting a Milestone in the Digital Transformation of Taiwan's Financial Security

In recent years, rampant fraud has become a major concern for social security. With the Chinese New Year's Eve approaching, scam gangs have begun to intensively lure the public through fraudulent messages, fake gift certificates, etc. In order to strengthen the prevention of fraud and various types of criminal transactions, Taipei Fubon Bank, following the "Eagle Eye Model," once again strikes to launch the newly-developed AI anti-money laundering "Owl System" prior to the Chinese New Year, which is the first anti-money laundering system with an AI intelligent model in the domestic banking industry that is capable of accurately detecting and verifying suspicious transactions, and quickly notifying and controlling them in order to block the flow of criminal funds and guard financial order and the safety of the public's assets in an all-rounded manner.

Kevin Tsai, SVP of Risk Management Division of Taipei Fubon Bank, said that "financial security" is the key to the sustainable operation of the financial industry and the promotion of social common good, as well as the focus of future financial digital development. In order to strengthen its core competence in financial security, Taipei Fubon Bank actively echoes Fubon Financial Holdings’ "Digital Transformation Blueprint for Financial Security" by applying diversified financial technologies in the areas of anti-fraud, anti-malpractice and anti-money laundering, as well as benchmarking against the international financial industry, and cooperating with global financial technology startups in the hope that through the introduction of forward-looking technologies, we can enhance the strength of financial security technology in Taiwan to accurately monitor and effectively combat illegal transactions that are harmful to society.

Kevin Tsai explained that the term "anti-money laundering" may be unfamiliar to the public, but in fact anti-money laundering is a very important part of crime deterrence. Taking fraud as an example, the ultimate goal of scam gangs is to get money without revealing their identities, so they often receive the proceeds of crime through dummy accounts. If suspicious accounts can be quickly and accurately identified, and notified and controlled in accordance with the laws and regulations of the competent authorities, it will be able to more effectively intercept unlawful behavior in the financial processing section, avoiding the transfer of fraud proceeds, and at the same time increase the difficulty for scam gangs to carry out their crimes, and inhibit the incidence of fraudulent cases.

However, in practice, banks have to handle a large number of financial transactions on a daily basis, so how to enhance the effectiveness and accuracy of monitoring so that criminal transactions cannot be hidden has also become an important issue in the enhancement of financial security. Taipei Fubon Bank has more than one million transactions that require initial monitoring every day, of which more than 6,000 transactions require further name verification, risk scoring or enhanced review. In the past, most of the key aspects of monitoring and verification were handled manually, which could not immediately reflect the changes in supervision regulations and the list database, and it also took a certain amount of time to conduct verification, making it difficult to stop suspicious fund flow in a timely manner.

At the same time, most banking systems adopt rule-based retrieval, and even if a few use AI technology, they still focus on auxiliary functions, and actual execution still relies on manpower. Therefore, the proportion of false alarms often remains high, which not only consumes a lot of system resources and processing manpower of the bank, but also requires customers to cooperate in explaining the purpose of funds, trading object, or providing transaction supporting documents which greatly affect the smoothness of transactions and service experience.

In view of this, since 2020, Taipei Fubon Bank has cooperated with Tookitaki, a leading brand of global anti-money laundering system, to develop the "Owl System", which is the first anti-money laundering system with AI intelligent model in the banking industry of Taiwan. It has a brand-new AI intelligent model comparable to the human brain, which is accurate in judgment and excellent in efficiency, and is able to automatically screen out high-risk transactions for follow-up processing and is able to process approximately 50 transactions per second, more than doubling the efficiency of the review process. In addition, the “Owl System” is equipped with self-learning capabilities which can automatically build machine learning models based on the latest risk management policy information of Taipei Fubon Bank, analyze customer behaviors and self-evolution from the audit results to improve the accuracy of abnormality monitoring, and it is estimated that it can reduce the number of false alarms by 45% at the initial stage of the system's launch.

With the mission of safeguarding the assets of customers and the public, Taipei Fubon Bank has led the industry in establishing the “Financial Security Department” since 2018 to coordinate the bank's anti-money laundering and fraud prevention and other financial security related businesses. Peiling Tsai, SVP of the Financial Security Department of Taipei Fubon Bank, said that the implementation of anti-money laundering work requires a high degree of professionalism and judgment, so the "Owl System" is not only a set of programs, but also is more similar to a new "AI colleague", with its own independent areas of responsibility, and at the same time, must continue to interact and exchange with other colleagues, and through the exchange of information to promote the growth of artificial intelligence and make the thinking and judgment more mature. The addition of this new colleague will allow the original anti-money laundering staff to focus on handling high risk alerts, and through the division of labor to assist in the overall strengthening of the effectiveness of crime prevention, which is not only a pioneering step in Taiwan's banking industry, but also a brand new milestone in the empowerment of financial technology.

With the wave of digitalization of financial services, various new and innovative products are flourishing, and criminal methods are changing day by day, resulting in the tightening of international anti-money laundering laws and regulations, which makes it more difficult for the banking industry to control risks. In order to protect the stability of the financial order, Taipei Fubon Bank took the lead in cooperating with the Criminal Investigation Bureau to develop the "Eagle Eye Model", which can detect suspected abnormal transfer behavior in advance and alert customers and employees to achieve the effect of preventing fraud beforehand. Under the guidance of the FSC and the National Police Agency, the "Eagle Eye Anti-Fraud Alliance" was launched and the Eagle Eye Model technology was provided to the members of the alliance free of charge, which was responded by 34 financial institutions, and the inter-institutional collaboration mechanism was established to share the information on fraud prevention. Now, combined with the newly launched "Owl System", it is expected to prevent fraudulent crimes from all aspects of the fund flow and provide the public with a more secure and fair trading environment.

Looking ahead, Taipei Fubon Bank will continue to pay attention to the development of financial technology and, with "Responsible Innovation" as the core, actively invest in technological innovation to drive the upgrading of financial services and the positive flow of fund so as to generate a stable and positive development force for the society, and to become the backing for the protection of the assets of the people in Taiwan.