On 15 October 2026, the Financial Stability Board published its final report on Sound Practices for Responsible Adoption of Artificial Intelligence, identifying 12 non-binding sound practices organized into two groups — organization-wide AI governance and risk management throughout the AI lifecycle — to help financial institutions adopt AI responsibly, with the report reflecting public feedback on a June 2026 consultative version.}
The same October 15 action finalized a list of 12 sound practices to guide AI adoption by banks and other financial institutions, framed as a practical toolkit that institutions can tailor to their context and integrate into existing cyber and third-party risk frameworks; the practices fall into organization-wide governance and AI-lifecycle risk management.
The board’s report explicitly labels the 12 practices non-binding, and they are meant to help financial institutions adopt AI responsibly. The practical toolkit described by the banking journal shares that intent, offering guidance institutions can shape to their own needs rather than a fixed template.
The two-group structure organizes the entire set. One group covers organization-wide AI governance; the other covers risk management throughout the AI lifecycle. Both the board’s final report and the banking journal’s account place the split along exactly those lines, confirming the practices span from institutional oversight to continuous lifecycle risk handling.
The finalized report did not appear in a single draft. It reflects public feedback on a consultative version published in June 2026, meaning the 12 practices absorbed an external comment period before reaching their final form. For a sector tracking AI guidance as deployments accelerate, that iterative path gives the output a basis in broader input.
The banking journal’s framing emphasizes tailoring. The practices are a toolkit institutions can adapt to their context, and the document envisions them being integrated into existing cyber and third-party risk frameworks rather than standing apart from current controls. That design lets a bank or other financial institution map AI considerations onto risk structures already in place.
Taken together, the FSB’s finalized output gives the sector a consolidated, non-binding reference of 12 practices across two groupings. The content is meant to be adopted responsibly by being adapted, not copied wholesale, and the lifecycle view of risk sits alongside organization-wide governance as the document’s structural core.