FSB finalises sound practices for the responsible adoption of artificial intelligence (AI)

While not binding in the UK, these global expectations will shape supervisory thinking and likely feed into national regulators' AI guidance, including potential FCA/PRA expectations. For compliance teams, the message is clear: AI governance must move from ad hoc experimentation to board-level accountability. Key areas include robust model risk management, fairness and bias testing, transparency and explainability, third-party and vendor oversight, data quality, and operational resilience.

Firms should conduct a gap analysis against the FSB practices, update AI policies and controls, and ensure senior management can evidence responsible AI decision-making. Actionable next steps: (1) map existing AI use cases and controls to the FSB practices; (2) strengthen AI risk assessments, validation and monitoring; (3) review contractual and diligence processes for AI vendors; (4) provide targeted board and staff training; and (5) engage with regulators on expectations. The opportunity is significant: firms that demonstrate responsible AI can accelerate innovation, improve efficiency and build trust, while those that lag face supervisory scrutiny, reputational damage and potential consumer harm.

RegCanary recommends treating this as a strategic compliance priority, not a one-off paper exercise.