AITOPAI热报
GlobalGuidelineIn forceFinanceDeployers

FSB Final Guidance on AI Risk Management

金融稳定理事会AI风险管理最终版指引
金融稳定理事会（FSB）·Published 2025-10-15·Checked by 管理员 on 2026-10-17

The Financial Stability Board (FSB) published its final guidance on AI risk management on 15 October, calling on banks and other financial institutions to strengthen human oversight of AI agents and define cases where agents must obtain human approval before acting. The final version strengthens recommendations on AI agents compared to the June draft and drops the suggestion to treat AI as "virtual employees".

Source ↗Remind meGlobal · All

What it means for you

  • Financial institutions must build human-approval mechanisms for high-risk AI uses, defining scenarios where agents need human sign-off
  • Boards must embed the guidance into AI risk governance and own oversight responsibilities
  • Fintech and AI vendors operating in global markets must align agent deployment and controls with FSB standards

    • What to do

    ComplianceMap current AI agent use cases and define high-risk steps requiring human approval
    LegalUpdate AI risk management policies and governance docs to reflect the final guidance
    EngineeringDesign human-in-the-loop approval and monitoring mechanisms for agent systems

    Key dates

    2025-10-15Published

    Source documents

    Source ↗

    Coverage

    Stories in the feed
    Editorial summary, not legal advice; dates and status follow the original document. · Report a correction