2024-06-23-国际清算银行-窥探炒作-评估高科技工具从实验到监督的转变（英）_20页_532kb

Executive Summary

The article examines the transition of supervisory technology (suptech) tools from experimentation to full integration in financial supervision. Suptech, defined as innovative technology such as AI and machine learning used to enhance supervisory work, has seen significant deployment but often fails to become truly critical due to implementation challenges. Key factors for success include a well-defined suptech strategy, user-friendly interfaces, access to granular data, and seamless integration into existing supervisory processes. While tools like data visualization and automation are commonly deployed, financial risk assessment requires overcoming technical and usability hurdles. The analysis, based on surveys of 32 supervisory authorities, underscores the need for a process-focused approach to maximize suptech's transformative potential, moving beyond issue-specific experiments to align with real-world supervisory challenges.

Main Findings on Deployment and Criticality

• Deployment Status: Despite high success rates in deploying suptech tools (around 90%), only about 53% of authorities report tools that are critical to supervision, indicating they are indispensable in key processes.
• Critical Factors:
  • Suptech Strategy: Authorities with a strategy are more likely to have critical tools, as strategies address deployment issues like rollout, integration, and continuous support.
  • User Accessibility: Tools must have direct, user-friendly interfaces and foster supervisor confidence to avoid barriers.
  • Data Availability: Granular data is essential for extracting insights and effectiveness; without it, deployment challenges arise.
  • Integration: Critical tools integrate naturally into the supervision IT ecosystem, allowing straight-through processing and reducing fragmentation.

Survey Results

• Deployment Success: Most authorities assess deployment success based on usage and time saved, with success rates high but actual criticality low.
• Areas of Focus: Data visualization, regulatory reporting, financial risk assessment, and automation lead in deployment; however, risk assessment tools face implementation difficulties.
• Global Participation: Involves 32 authorities from diverse jurisdictions, highlighting converging needs but regional variations in resources.

Challenges in Effective Embedding

• Usability Issues: Many tools require specialized skills or lack direct interfaces, limiting adoption.
• Technical Constraints: Data access problems, on-premises computing power, and cloud hesitations hinder effectiveness.
• AI-Specific Problems: GenAI tools are use-case specific and need careful expectation management to avoid unmet promises.

Conclusion and Recommendations

• Suptech should enhance efficiency and effectiveness by addressing supervisor needs and embedding seamlessly, rather than creating new processes. The analysis calls for international efforts to focus on process rather than specific issues, improving replicability and impact. Supervisory authorities are urged to prioritize long-term IT ecosystem development, data quality, and skill enhancement to fully leverage suptech's potential. This ensures suptech moves beyond hype to become a core workhorse for financial stability.