The Financial Stability Board (FSB), the international body that coordinates financial regulators from the world’s major economies, has published the final version of its sound practices for the responsible adoption of artificial intelligence. The 12 practices are guidance, not rules: the FSB describes them as “a practical toolkit that financial institutions can tailor to their specific context and risk profile.”
The FSB announced the final report on Thursday, 15 October 2026. It follows a consultation on a draft version published in June 2026, and the FSB has also released an overview of the responses it received.
What the report says
The report, Sound Practices for Responsible Adoption of Artificial Intelligence, sets out 12 practices meant to help financial institutions “harness the benefits of AI while managing its risks responsibly, in ways that support the resilience of the financial system.” They fall into two groups.
The first group covers organisation-wide AI governance. According to the FSB, it highlights:
- (i) the board and senior management’s “pivotal role” in making sure AI adoption fits the firm’s business model, risk appetite and strategy;
- (ii) a governance framework with well-defined roles and responsibilities;
- (iii) building AI risks into the risk management framework, with effective documentation of how AI is used; and
- (iv) continuous learning and adaptation.
The second group sets out a framework for managing risk across the AI lifecycle, from development to deployment:
- (v) assessing the materiality and risks of each AI use case at the start and afterwards;
- (vi) choosing appropriate AI technologies;
- (vii) appropriate data governance;
- (viii) explainability of different types of AI, and compensating controls where needed;
- (ix) evaluating how AI use cases perform;
- (x) appropriate and effective human oversight;
- (xi) managing AI-related cyber and information and communication technology (ICT) risks; and
- (xii) managing risks from third-party AI dependencies.
How firms are expected to use it
The FSB is not imposing requirements. Instead, it says it “strongly encourages the board and senior management of financial institutions to reference the sound practices when developing and refining their institution’s AI governance, risk management and control frameworks.”
Firms can fold the practices into frameworks they already have, including those for cyber and third-party risk, the FSB says. The report includes case studies showing benefits that financial institutions have reported and how some practices have been applied in proportion to the size and risk of a firm’s activities.
The practices also aim to improve coordination and information-sharing between financial institutions, supervisors and others, both within countries and across borders. The FSB says they build on, and are broadly compatible with, existing work by the FSB and other standard-setting bodies.
Why it matters
The FSB brings together national authorities responsible for financial stability in 24 countries and jurisdictions, along with international financial institutions and groupings of regulators. It also reaches about 70 other jurisdictions through six Regional Consultative Groups. Its guidance carries weight because national supervisors who sit on the board often use it when shaping their own expectations, even when it has no legal force.
For banks, insurers and asset managers, the list reads as a checklist of what supervisors are likely to ask about: who on the board owns AI risk, how each use case is assessed, whether models can be explained, how humans stay in the loop, and what happens if a key outside AI provider fails. The final two items, on AI-related cyber risk and third-party dependencies, reflect themes the FSB and other bodies have been raising this year.
The FSB is chaired by Andrew Bailey, Governor of the Bank of England. Its Secretariat is in Basel, Switzerland, hosted by the Bank for International Settlements.
What’s still open
TSN has read the FSB’s press release but not yet the full final report or the overview of consultation responses, so we are not reporting what changed between the June draft and the final version. The FSB says it will monitor how AI technologies and applications evolve and “update the sound practices as necessary.”
What to watch
Watch whether national supervisors refer to the FSB practices in their own AI guidance, and how quickly. Regulators in several markets are already publishing their own expectations; on the same day, the Hong Kong Monetary Authority and New Zealand’s regulators set out AI-related guidance of their own.
What this means
- Guidance, not law. The 12 practices are non-binding, but boards are strongly encouraged to use them.
- Governance first. Half the framing is about board responsibility and clear accountability.
- Cyber and suppliers in focus. AI-related cyber risk and third-party AI dependencies are named practices.
Sources
- Financial Stability Board, “FSB finalises sound practices for the responsible adoption of artificial intelligence (AI)”, press release Ref 14/2026, 15 October 2026. https://www.fsb.org/2026/10/fsb-finalises-sound-practices-for-the-responsible-adoption-of-artificial-intelligence-ai/
- Financial Stability Board, Sound Practices for Responsible Adoption of Artificial Intelligence (AI): Final report, 15 October 2026. https://www.fsb.org/2026/10/sound-practices-for-responsible-adoption-of-artificial-intelligence-ai-final-report/
