Can AI Trigger the Next Financial Crisis? Algorithmic Trading Black Boxes and the Responsibility Gap

With the rapid advancement of artificial intelligence technologies, algorithm-based investment systems have been spreading rapidly in financial markets. High-frequency trading, automated portfolio management, and machine-learning-based investment strategies are fundamentally transforming the structure of investment decision-making. This transformation goes beyond simple technological innovation and is reshaping both the judgment structure and the responsibility structure of financial markets. 1 This study begins with the following questions: Can AI-driven investment systems become the trigger of the next financial crisis? If such a situation occurs, who should be held responsible? This paper introduces the concept of a structural responsibility gap that emerges in algorithmic investment environments. In automated investment systems, data col lection, model analysis, signal generation, and order execution operate within a unified algorithmic structure. Investment decisions are therefore performed internally within the system without direct human intervention, while responsibility becomes distributed across developers, financial institutions, investors, and regulatory authorities. As a re sult, a structural problem arises in which the attribution of responsibility becomes unclear in times of crisis. Thestudy also analyzes how algorithmic synchronization and feedback loops—which occur when multiple algorithmic trading systems simultaneously respond to shared market signals—can amplify market volatility and expand systemic risk. Such dynam ics provide an important mechanism for explaining sudden market instabilities such as f lash crashes. To address this issue, the paper proposes a Hierarchical Joint Responsibility Model. This model restructures responsibility within algorithmic investment environments into three layers: the design layer, the operational layer, and the authorization layer. By doing so, it presents an institutional framework capable of maintaining both traceability and normative control of responsibility within automated financial systems. This study emphasizes that AI-based financial systems are not merely technologi cal tools but socio-technical structures that transform the very architecture of finan cial judgment. The study ultimately aims to provide philosophical and institutional foundations for redesigning responsibility structures in increasingly automated finan cial markets, while also offering conceptual guidance for regulators and policy makers seeking to address accountability challenges in AI-driven financial systems.