FSB Revises Financial AI Guidance With Focus on Autonomous Systems

The Financial Stability Board’s report, Sound Practices for Financial Institutions’ Responsible AI Adoption: Overview of Consultation Responses, clarifies oversight for autonomous AI and outside vendors while leaving financial-system-wide risks beyond the guidance.

The consultation ran from June 10 to July 22, 2026, and drew around 160 responses from financial firms, technology providers, academics, public authorities and other stakeholders. Participants broadly supported practices that firms can adapt to their size and risk profile and apply across technologies. Some called for more specific guidance.

The consultation surfaced concerns about evaluating external AI models, especially frontier models, when vendors provide too little information for due diligence. It also highlighted financial institutions’ reliance on key providers.

For agentic AI, participants called for safeguards including approval before systems act, limits on access, audit trails, real-time monitoring, clear accountability and ways for people to intervene.

The final guidance clarifies that documentation should support accountability and help firms trace and reconstruct actions, particularly in agentic AI use cases. It adds a case study on agentic AI and examples involving insurers and other nonbank firms. The final report modifies Sound Practice 12 to add guidance on managing risks from third-party AI providers.

The practices remain proportionate and technology-neutral guidance, rather than an international standard prescribing particular AI choices. The guidance did not add measures for financial-system-wide or macroprudential risks, including correlated behavior among firms using shared AI models. The FSB will continue monitoring AI-related systemic risks through regular vulnerability assessments. Its practices for agentic and multi-agent systems remain under development and may evolve as understanding of their risks grows.