인공지능을 활용한 금융감독의 법적 함의

Financial supervisory authorities around the world are increasingly using artificial intelligence (AI), machine learning (ML), and innovative strategies that involve big data analytics for digital financial supervision. Traditionally, financial authorities relied primarily on information from the regulatory reports of financial companies and securities issuers for their supervisory and market monitoring activities. More recently, AI analytics have enabled authorities to expand their market monitoring activities and to analyze vast amounts of unstructured data culled from non-traditional information sources, such as online news outlets, social media, blogs, forums, and consumer complaints. Financial authorities have also deployed AI tools to enhance the efficiency and effectiveness of traditional market monitoring tasks to detect misconduct, fraud, and illegal behavior, and to protect consumers. The US Securities Exchange Commission (SEC) uses AI to target high-risk activities and potential actors, thus improving its ability to allocate scarce agency resources. The relatively slow development of the Korean market monitoring authorities – which include the Financial Supervisory Service (FSS) and the Korea Exchange (KRX) – reflects the challenges of implementing an effective AI strategy and the extent of the resources necessary for doing so. Authorities should ensure the standardization, accuracy, completeness, and integrity of the data, and must address the algorithmic bias issues. A government’s use of AI, especially in areas of financial supervision and enforcement, raises urgent issues with regard to rule of law in administrative actions, accountability, and the transparency of financial supervision. In conjunction with the development of AI, legislators’ and policy makers need to review their country’s financial, data protection, and administrative laws. Given the black box-like nature of AI, policymakers should consider the tension between accountability and the (un)explainability of AI.