Financial Stability Board issues AI governance guidance for banks

The Financial Stability Board (FSB) has issued final guidance for financial institutions regarding the adoption of artificial intelligence, specifically focusing on “agentic” AI. These AI agents are capable of planning and executing tasks with minimal human intervention, such as moving data or making autonomous decisions within company systems.

The FSB’s framework includes 12 sound practices divided into two main categories: organization-wide AI governance and a comprehensive framework for managing risks throughout the AI lifecycle. The guidance emphasizes that senior management and boards must oversee the alignment of AI adoption with business models and risk appetite.

Notably, the FSB moved away from previous suggestions to treat AI agents as “synthetic employees,” warning that these systems do not reason like humans and may interpret goals in unexpected ways. The guidance calls for increased human oversight for high-risk tasks and the establishment of clear, risk-based triggers that require human approval before an AI agent proceeds with certain actions.