The Financial Stability Board (FSB) has developed a comprehensive set of sound practices designed to guide all types of financial institutions in responsibly navigating the significant benefits and inherent risks associated with the increasing adoption of Artificial Intelligence (AI). This consultation report aims to establish a robust framework for managing these evolving challenges and opportunities within the financial sector.
Financial institutions are actively leveraging Artificial Intelligence to fundamentally transform their operational efficiencies and enhance service delivery. This rapid integration, however, also introduces or potentially amplifies various risks that necessitate careful identification and proactive management. The report emphasizes that a responsible approach to AI adoption enables these institutions to fully capitalize on new opportunities and benefits while concurrently mitigating the associated perils. Crucially, institutions must develop a deep understanding of AI's evolving landscape, stay informed about the latest developments, and implement appropriate strategic adoption frameworks and protective measures to manage these dynamic risks effectively. From a broader financial system perspective, promoting responsible AI adoption is paramount to safeguarding overall financial stability.
To foster and facilitate the responsible integration of AI across the financial sector, this report introduces a detailed menu of twelve sound practices. These practices are designed to be applied by financial institutions at an organization-wide level, covering various aspects of AI governance and the comprehensive management of the AI development and deployment lifecycle. The framework provided includes practical case studies, derived from actual AI implementation scenarios within financial institutions. These real-world examples serve to illustrate how the proposed sound practices can be effectively applied in practical settings and, importantly, how their implementation can be scaled and adapted proportionately to suit different institutional contexts and AI use cases. The overarching goal of these sound practices is to assist the board and senior management teams of financial institutions in making informed decisions regarding their business strategy, technology adoption roadmaps, and risk management protocols within the rapidly evolving, AI-enabled operational environment. This initiative builds upon extensive existing and ongoing work conducted by the FSB itself, other pertinent standard-setting bodies, and national and regional financial authorities globally. Furthermore, it integrates valuable insights gathered from a diverse array of stakeholders across the financial ecosystem, including various financial institutions and their technology vendor partners, ensuring a holistic and practical approach.
The Financial Stability Board is actively seeking detailed feedback on this consultation report through a series of critical questions posed to stakeholders. These questions are structured to gather comprehensive perspectives on several key aspects. Firstly, the FSB asks if respondents concur with the outlined benefits and risks of AI adoption by financial institutions, and if there are any significant advantages or dangers that have not been adequately covered. Secondly, it inquires whether the proposed sound practices are sufficiently comprehensive and clear to effectively guide financial institutions in their responsible AI adoption efforts. Thirdly, the consultation probes whether the practices achieve an appropriate balance in managing general AI risks while also specifically addressing the unique challenges posed by new and complex forms of AI, such as Generative AI (GenAI) and agentic AI. Fourthly, it seeks opinions on the flexibility of these sound practices to adapt and remain relevant as newer types of AI technologies emerge and as responsible adoption methodologies evolve over time. Fifthly, the FSB requests feedback on whether the included case studies adequately demonstrate how various financial institutions can benefit from responsible AI adoption and if additional case studies, particularly from non-bank financial entities, should be incorporated. Sixthly, it questions if the case studies offer actionable insights for institutions in their practical implementation of responsible AI. Finally, the consultation asks if the definitions provided in the glossary are clear, precise, and align effectively with current industry sound practices and recent advancements in AI technology.