Manage Risk with the Financial Stability Board’s 12 Sound Practices for Responsible AI Adoption Framework

The Financial Stability Board’s (FSB’s) 12 sound practices report offers a framework for managing artificial intelligence (AI) risks. The report brings together emerging approaches to responsible AI governance in financial institutions, with potential implications for multinational institutions operating across jurisdictions, writes Skadden in a recent article. 

Published in June 2026, the FSB’s Sound Practices for Responsible Adoption of Artificial Intelligence (AI) does not establish new international standards. Instead, it consolidates emerging governance, accountability, documentation, and life-cycle risk management practices. It draws on case studies from banks and insurers that demonstrate how organizations are deploying AI, addressing governance challenges, and developing controls as adoption expands. 

The framework’s significance lies in reinforcing the need to incorporate AI governance into existing governance, operational risk, and compliance frameworks rather than treating it as a separate technology initiative.

The report’s recommendations align with the European Union (EU) AI Act, the Digital Operational Resilience Act (DORA), and U.S. financial regulatory expectations. For multinational financial institutions, the FSB framework offers a practical operating model that can serve as a common global baseline while still allowing for jurisdiction-specific legal requirements.

Beyond governance, the FSB’s 12 sound practices address strategic direction, accountability, risk management, model selection, transparency, performance monitoring, cyber resilience, and third-party risk. The practices emphasize a life-cycle approach in which organizations maintain inventories, document decisions, assess risks proportionately, and continuously test and monitor AI systems.

Case studies in the report highlight recurring challenges. For example, third-party AI can create information and accountability gaps, while expanding adoption can expose limitations in decentralized governance. 

Financial institutions that have a mature AI governance framework can build on it. For those that do not, centralized AI inventories, stronger vendor oversight, and effective human review can provide practical starting points.