有效運用人工智慧打擊洗錢犯罪之研究

Research on the Effective Use of Artificial Intelligence to Combat Money Laundering Crimes

10.6868/HKLR.202406_(76).05

楊暹黼(Hsuan-Fu Yang)

洗錢 ； 洗錢防制 ； 科技管制 ； 人工智慧 ； 資料在地化 ； 數位主權 ； Money Laundering ； Anti-Money Laundering ； Technological Regulation ； Artificial Intelligence ； Data Localization ； Digital Sovereignty

華岡法粹

76期（2024 / 06 / 01）

147 - 202

繁體中文;英文

洗錢防制，在洗錢手段日益複雜、資恐金源五花八門，以及科技洗錢手段遍地開花之年代，逐漸成為法律上、政策上之顯學。本文首先地毯式介紹當代洗錢手段之複雜性，並討論晚近透過虛擬貨幣等科技手段，進行不法交易或洗錢之具體情形，從而導引至當代反洗錢法制發展與整合之脈絡。其次，基於洗錢日漸複雜之趨勢，以及人力反洗錢稽核程序之效率低落，論述引進人工智慧防制洗錢之實益，並介紹晚近反洗錢演算法之發展趨勢，以及市面上人工智慧反洗錢方案之推陳出新。再者，新科技之應用絕非一帆風順，本文接續討論人工智慧反洗錢方案可能面臨之困難，其中包括各國尚在發展階段之人工智慧政策差異、資料在地化趨勢對反洗錢系統資料庫整合帶來之障礙，以及數位主權爭議對反洗錢演算法可能帶來之衝擊。

Within the realm of countering money laundering, a landscape characterized by progressively intricate methods, diverse sources of terrorism financing, and the widespread proliferation of technology-driven illicit activities, it has undeniably ascended to a pivotal position in legal and policy discourses. This paper embarks on an exhaustive exploration, commencing with a detailed exposition of the intricacies inherent in contemporary money laundering techniques. It meticulously scrutinizes recent instances wherein illicit transactions or money laundering activities have adeptly exploited advanced technological means, such as virtual currencies, prompting an in-depth inquiry into the evolving landscape of contemporary anti-money laundering legal frameworks. In response to the mounting intricacies of money laundering and the palpable inefficiencies associated with manual anti-money laundering audit procedures, this paper ardently advocates for the manifold benefits arising from the strategic integration of artificial intelligence into money laundering prevention methodologies. Beyond introducing recent trends in developing anti-money laundering algorithms, it sheds light on the continuous innovation within artificial intelligence solutions designed explicitly to counteract the multifaceted challenges of evolving money laundering activities. Moreover, the paper strategically pivots towards a nuanced discussion surrounding the potential challenges artificial intelligence anti-money laundering solutions might face shortly. These challenges encompass the variations in artificial intelligence policies among nations still navigating developmental stages and the hindrances arising from the overarching trend towards localized data. Additionally, the paper carefully addresses the potential impacts on anti-money laundering algorithms stemming from disputes related to digital sovereignty, further complicating the landscape. As the international community grapples with these multifaceted challenges, this paper seeks to enrich the ongoing discourse by providing profound insights into the intricate nature of money laundering prevention in an era characterized by rapid technological advancements and the ever-expanding web of global financial interconnectivity.