The Financial Stability Board published its consultation report 'Sound Practices for Responsible Adoption of Artificial Intelligence' on 10 June 2026. The report identifies 12 practices to assist financial institutions of all types in managing the benefits and risks of AI adoption at the organisational and use-case level. The report is at the consultation stage: the FSB invites written responses by 22 July 2026 and will publish a final report in October 2026. The 12 practices are non-binding guidance and do not constitute FSB standards.
Practices 1 through 4 address organisation-wide AI governance, covering the internal structures and processes that should govern decisions about whether and how to adopt AI at scale. Practices 5 through 10 address lifecycle management of individual AI use cases, requiring that deployments be supported by proportionate guardrails at each development and deployment stage. Practices 11 and 12 address ongoing monitoring and escalation. The FSB grounds the report in existing guidance from the Basel Committee on Banking Supervision, IOSCO, the International Association of Insurance Supervisors, and G20 member regulators, without creating binding standards of its own.
Banks, investment managers, insurers, and systemically important financial institutions are the primary audience. The FSB explicitly acknowledges that continuous human monitoring of individual AI agent decisions becomes impractical as agentic AI systems multiply inside financial institutions, and recommends supplementing human oversight with AI systems that monitor other AI systems. Compliance teams, model risk officers, and AI governance functions at financial institutions should review the consultation text and assess whether current governance arrangements meet the 12 practices before responding.
The consultation is a non-binding soft-law document, and the October 2026 final report will carry the same status. National regulators are not required to implement the FSB's sound practices into binding prudential rules or supervisory expectations. Whether member jurisdictions incorporate the practices into enforceable guidance will depend on each regulator's own legislative programme. The FSB's acceptance of AI-monitoring-AI architectures as legitimate risk management signals that future binding guidance from national regulators may adopt similar positions.
Licentium advises financial institutions and their counsel on AI governance requirements across EU, UK, and international regulatory regimes. Our team and partner network are available to assist with FSB sound practice gap analyses, model risk management policy updates, and regulatory consultation responses. Work we undertake includes AI governance reviews for regulated financial entities, prudential compliance assessments for AI systems, regulatory consultation submissions, and cross-border AI deployment advice for banks, asset managers, and insurers.