Taiwan's Financial Sector Strengthens Anti-Fraud Efforts Through Cross-Domain Collaboration, Enhancing AI Model Capabilities
To combat increasingly sophisticated fraud, Taiwan's financial sector is enhancing its defenses through cross-domain collaborations. On the 21st, Cathay United Bank announced a deeper partnership with the Criminal Investigation Bureau to implement an AI model for analyzing victim behavior, which is expected to intercept NT$90 million in fraudulent funds annually. This collaboration uses police data to improve model accuracy. Meanwhile, Fubon Insurance has partnered with Taiwan Mobile and Gogolook to launch a digital anti-fraud picture book to raise public awareness.
📋 Article Processing Timeline
- 📰 Published: May 21, 2026 at 13:30
- 🔍 Collected: May 21, 2026 at 14:01 (31 min after Published)
- 🤖 AI Analyzed: May 21, 2026 at 14:06 (4 min after Collected)
(CNA, Taipei, May 21, by reporter Su Su-yun) Facing rapidly evolving fraud techniques, the financial industry is continuously upgrading its technological anti-fraud capabilities. Cathay United Bank, in partnership with the Criminal Investigation Bureau (CIB), has introduced a behavioral analysis model for victim typologies, estimated to intercept up to NT$90 million in fraudulent funds annually. Fubon Insurance has also teamed up with Taiwan Mobile and Gogolook to launch a new digital anti-fraud picture book.
Cathay United Bank announced today that it is deepening its cooperation with the CIB of the National Police Agency, Ministry of the Interior. This year, for the first time, the two parties have extended their anti-fraud mechanism to include behavioral analysis of 'victim typologies.' They have constructed an AI model training and validation mechanism that complies with regulations and protects personal data privacy, incorporating data and patterns from fraud cases investigated by the CIB to enhance the ability of front-line bank staff to identify abnormal cash flows and behaviors in real-time.
Cathay United pointed out that a single financial institution relying solely on its internal data can hardly grasp the full behavioral characteristics of victims, leading to detection blind spots. To close this security gap, the two parties initiated data cooperation starting from 2025, combining fraud cases known to the CIB, including changes in account usage behavior and digital channel fraud patterns. This data is used for model validation, induction, and updates, significantly improving overall detection accuracy and timeliness.
Cathay United Bank preliminarily estimates that it can intercept up to NT$90 million in fraudulent funds annually. Looking ahead, Cathay United will continue to deepen this joint defense mechanism and expand cross-domain cooperation in compliance with legal regulations and information security governance principles. The bank will also enhance its AI models to strengthen the identification of subtle and diverse victim behaviors.
Fubon Insurance has also launched a new digital anti-fraud picture book, "Phew's Lab - Financial Anti-Fraud Classroom: The Anti-Fraud Helper Edition," in collaboration with Taiwan Mobile and Gogolook. The picture book presents professional anti-fraud knowledge and methods to strengthen the public's financial anti-fraud concepts and awareness.
Fubon Insurance stated that with fraud methods constantly changing, there is a learning need for everyone, including the general public, children, and parents, regarding fraud patterns and risk identification. The newly launched anti-fraud picture book uses a story scenario of encountering fraud to guide people to recognize common fraudulent calls, text messages, and websites. Through character interaction and cases close to daily life, it aims to help them grasp key anti-fraud points, cultivate the ability to identify risks, and thereby reduce the probability of being scammed.
Cathay United Bank announced today that it is deepening its cooperation with the CIB of the National Police Agency, Ministry of the Interior. This year, for the first time, the two parties have extended their anti-fraud mechanism to include behavioral analysis of 'victim typologies.' They have constructed an AI model training and validation mechanism that complies with regulations and protects personal data privacy, incorporating data and patterns from fraud cases investigated by the CIB to enhance the ability of front-line bank staff to identify abnormal cash flows and behaviors in real-time.
Cathay United pointed out that a single financial institution relying solely on its internal data can hardly grasp the full behavioral characteristics of victims, leading to detection blind spots. To close this security gap, the two parties initiated data cooperation starting from 2025, combining fraud cases known to the CIB, including changes in account usage behavior and digital channel fraud patterns. This data is used for model validation, induction, and updates, significantly improving overall detection accuracy and timeliness.
Cathay United Bank preliminarily estimates that it can intercept up to NT$90 million in fraudulent funds annually. Looking ahead, Cathay United will continue to deepen this joint defense mechanism and expand cross-domain cooperation in compliance with legal regulations and information security governance principles. The bank will also enhance its AI models to strengthen the identification of subtle and diverse victim behaviors.
Fubon Insurance has also launched a new digital anti-fraud picture book, "Phew's Lab - Financial Anti-Fraud Classroom: The Anti-Fraud Helper Edition," in collaboration with Taiwan Mobile and Gogolook. The picture book presents professional anti-fraud knowledge and methods to strengthen the public's financial anti-fraud concepts and awareness.
Fubon Insurance stated that with fraud methods constantly changing, there is a learning need for everyone, including the general public, children, and parents, regarding fraud patterns and risk identification. The newly launched anti-fraud picture book uses a story scenario of encountering fraud to guide people to recognize common fraudulent calls, text messages, and websites. Through character interaction and cases close to daily life, it aims to help them grasp key anti-fraud points, cultivate the ability to identify risks, and thereby reduce the probability of being scammed.
FAQ
國泰世華銀行如何強化防詐?
該行與刑事局合作,導入「被害者樣態」AI行為分析模型,利用警方的詐騙案件數據來訓練模型,以提升行員辨識異常金流的能力。
國泰世華與刑事局的合作預計能達到什麼成效?
國泰世華銀行初步估計,該合作每年可望攔阻詐騙金額達新台幣9000萬元。
富邦產險採取了哪些防詐措施?
富邦產險攜手台灣大哥大和Gogolook,推出名為「萬事保研究所-金融防詐小教室:防詐邦手篇」的數位繪本,旨在透過故事情境教育民眾,提升其防詐觀念。
為何金融機構需要與警方跨域合作防詐?
因為單一金融機構僅靠內部資料,難以完整掌握被害者行為特徵,容易產生偵測盲點。結合警方的受詐案例數據能補強防護缺口,顯著提升偵測精準度與即時性。
這則新聞提到了哪些公司與機構的合作?
新聞中提及的合作案包含:國泰世華銀行與內政部警政署刑事警察局;以及富邦產險、台灣大哥大、走著瞧(Gogolook)的合作。