What happened
The Financial Stability Board released a 66-page consultation report on 10 June 2026 proposing 12 sound practices to guide all types of financial institutions in the responsible adoption of AI. The practices are organised across three pillars: (1) organisation-wide AI governance (practices 1–4); (2) management and mitigation of AI risks throughout the development and deployment lifecycle (practices 5–10); and (3) management of AI-related cyber, ICT, and third-party risks (practices 11–12). The report includes case studies drawn from real-world AI implementation at financial institutions and is explicitly calibrated to be compatible with existing work from the Basel Committee, IOSCO, and IAIS. Consultation closes 22 July 2026, with the FSB inviting comment on whether the practices are sufficiently flexible to address future AI types.
Why it matters
This is the first G20-backed, cross-sector global framework for AI governance in finance, giving CROs, CISOs, and compliance teams at banks, insurers, and asset managers the authoritative reference point against which supervisors will benchmark AI programmes. Boards at systemically important institutions should expect supervisory questions framed around these 12 practices within 12–18 months.
Action needed
Map your institution's existing AI governance programme against the 12 FSB sound practices and identify gaps before the July 22 consultation deadline; consider submitting comments on practices that would impose disproportionate compliance costs on your business model.