理財機器人之資訊揭露法制——以演算法之說明義務為中心

Abstract

理財機器人以演算法為基礎提供投資建議，雖提升金融服務之普及性與可近用性，然其自動化特性亦使資訊揭露與說明義務之落實面臨挑戰。金融商品本身即具有高度複雜性，而理財機器人更以演算法為核心提供投資顧問服務，要求業者說明其運作機制與投資建議產出方法等關鍵資訊，如何在兼顧技術特性下提升資訊揭露之有效性，使金融消費者能充分理解，成為實務上更具困難之課題。 行為經濟學之有限理性概念指出，單純的資訊揭露並不足以確保金融消費者受到充分保護，因其行為偏誤將影響理性自主決策之能力。因此，須透過妥適的揭露制度設計，方能有效實現說明義務之規範目的。就理財機器人而言，金融消費者真正需要理解的並非演算法本身，而在於該系統運用哪些參數、如何產出投資建議，及其相關限制與潛在風險。 基此，本文以理財機器人之資訊揭露與說明義務為核心，探討自動化投資顧問服務於金融消費者保護法制下之適用與挑戰。透過行為經濟學分析之應用，並參酌英美等國之比較法制經驗，本文提出演算法說明義務於內容與呈現方式上之具體改善方向，以降低金融消費者之行為偏誤並促進其對資訊之充分理解；進而於立法論與解釋論層面，提出實踐揭露與說明義務之可行方案，以兼顧金融消費者保護與市場競爭之平衡。

Robo-advisors, which provide investment advice based on algorithms, have enhanced the accessibility and inclusiveness of financial services. However, their automated nature poses significant challenges for the implementation of information disclosure and explanatory duties. Financial products are inherently complex, and when investment advice is generated by algorithms, requiring providers to explain operational mechanisms and the logic behind investment recommendations raises practical difficulties. Ensuring the effectiveness of such disclosures, while allowing consumers to fully understand the services, remains a critical issue. The concept of bounded rationality in behavioral economics indicates that information disclosure alone is insufficient to protect consumers, as behavioral biases impair their ability to make rational and informed decisions. Therefore, an appropriately designed disclosure framework is essential to fulfill the normative purpose of explanatory duties. With respect to Robo-advisors, what consumers need to understand is not the algorithmic code itself, but rather the parameters on which the system relies, the manner in which investment recommendations are generated, and the inherent limitations and potential risks of such automated advice. Accordingly, this thesis focuses on the information disclosure and explanatory obligations of Robo-advisors, examining their application and the challenges they pose within the financial consumer protection regime. Drawing on insights from behavioral economics and the regulatory experiences of the United States and the United Kingdom, this thesis proposes improvements to the content and presentation of algorithmic disclosures to mitigate consumer biases and promote informed understanding. It further suggests feasible approaches to implementing disclosure and explanatory duties at both the legislative and interpretive levels, balancing consumer protection with market competition.