Public consultation · International body
Sound Practices for Responsible Adoption of Artificial Intelligence (AI): consultation report
The FSB offers a menu of 12 practices for adopting AI responsibly, proportionate to the materiality and risk of each use.
- Organization
- Financial Stability Board (FSB)
- Document date
- 10 June 2026
- Type
- Public consultation
- Language
- English
- Scope
- International
- Length
- 66 pages
What it contains
The consultation report, published on 10 June 2026, identifies 12 sound practices to help all types of financial institutions manage the benefits and risks of adopting artificial intelligence. It starts from the fact that institutions already use AI in credit risk assessment, trading, portfolio optimisation, fraud prevention, transaction monitoring, customer service, coding assistance and document review, and that generative and agentic AI have attracted growing interest. The practices are not intended to set an international standard or a prescriptive approach, are not exhaustive and are expected to be refined as the technology evolves.
Practices 1 to 4 concern organisation-wide governance: strategic direction and oversight by the board and senior management, roles and responsibilities, incorporation of AI risks into the risk management framework and organisational adaptability. Practices 5 to 10 cover the lifecycle of each use case: materiality and risk assessment, selection of models or systems, data governance, explainability and transparency, performance management and human oversight. Practices 11 and 12 address AI-related cyber and information and communication technology risk, and the risk of third parties that provide AI, focusing on performance, transparency, data quality, supply chain, concentration and business continuity.
The report insists on proportionality: more robust practices for large, complex and interconnected institutions or where AI is used in critical functions, and only the relevant ones, or adapted ones, for smaller institutions or lower-risk uses. It includes an overview of adoption, use cases and benefits, an analysis of risks and implementation challenges, annexes with examples of explainability and performance testing, and a glossary.
Why it matters for risk management
For those who manage model, technology and third-party risk, it is an orderly framework for reviewing internal policies: an inventory of AI uses, materiality criteria, explainability required case by case and dependence on providers. Because it is a consultation version and non-binding, it should be treated as a reference alongside local regulation.
This summary describes the public consultation version (comments were received until 22 July 2026). The text may change in the final report; the summary will be updated when it is published.
Summary prepared by the Dominican Republic Risk Management Club from a reading of the original document, with the source verified on 12 October 2026. It does not replace reading the official text and is not advice.
Connect with our networks