Supporting Adoption of Artificial Intelligence in Fighting Financial Crime (2026-06-22)

Circulars Email: HKMA E-mail Alert of 23 June 2026 (05:00 p.m. HKT)

Document Information

Title: Supporting Adoption of Artificial Intelligence in Fighting Financial Crime (2026-06-22)

Type: Circulars

URL: https://brdr.hkma.gov.hk/eng/doc-ldg/current/20260622-1-EN

Email Received: 2026-06-23 19:12

Summary Created: 2026-06-23 13:19

English Summary
Quick section switch
Management Summary
  • Purpose / Background: The HKMA is accelerating the adoption of Artificial Intelligence (AI) by Authorized Institutions (AIs) to enhance the detection and mitigation of Money Laundering (ML) and Terrorist Financing (TF) risks. This circular introduces a new report summarizing AI deployment strategies, successful use cases, and the transition toward holistic, end-to-end AML/CFT operating models.
  • One-line conclusion: AIs must maintain their previously submitted AI implementation plans, continue exploring advanced AI use cases, and align internal systems with industry best practices outlined in the HKMA report.
  • Key Changes:
  • Shift from manual/legacy monitoring to AI-driven, holistic suspicious activity monitoring.
  • Integration of advanced use cases: dynamic risk assessment, facial watch lists, and corporate mule detection.
  • Focus on "Agentic AI" as the next evolutionary step in AML/CFT systems.
  • Enhanced expectations for AIs to treat their AML/CFT operating models as end-to-end AI-transformed systems.
  • Requirement to keep AI implementation plans updated for future HKMA review.
  • Key Dates / Deadlines:
  • 23 June 2026: Second workshop on "Agentic AI" at HKUST.
  • Ongoing: Ad-hoc reporting of progress and new AI use cases upon HKMA request.
  • Applicability / Impact scope: All Authorized Institutions (AIs) in Hong Kong, with specific emphasis on those with significant operations that have already submitted implementation plans.
  • Recommended management actions:
  • Review the "Supporting Adoption of Artificial Intelligence in Fighting Financial Crime" report to benchmark internal AI capabilities.
  • Assess current AI implementation plans against the provided use cases (e.g., mule detection, dynamic risk assessment).
  • Prepare for future regulatory requests for updates on AI deployment progress.
  • Evaluate the potential integration of "Agentic AI" into existing AML/CFT frameworks.
  • Ensure regular communication with the HKMA AML and Financial Crime Risk Division regarding technological transitions.
Detailed Summary

1) Document overview

  • Nature: Official circular from the HKMA.
  • Purpose: To provide guidance and share industry findings on the deployment of AI in AML/CFT, following the HKMA’s commitment to modernizing financial crime monitoring.
  • Scope: Applies to all Authorized Institutions (AIs).

2) Main requirements

  • AIs are expected to continue the deployment of AI to strengthen AML/CFT efficiency.
  • AIs that have submitted AI implementation plans must maintain these documents as "living" records.
  • Institutions should actively explore "holistic approaches" to suspicious activity monitoring as part of broader digital transformation strategies.

3) Key changes

  • Increased focus on "Agentic AI" capabilities.
  • Move toward end-to-end transformation of operating models rather than siloed AI implementation.
  • Standardized recognition of dynamic risk assessment and corporate mule detection as essential AI-driven use cases.

4) Important dates & transition

  • 23 June 2026: Knowledge-sharing workshop on Agentic AI (HKUST).
  • Ongoing: AIs must be ready to report progress/updates when formally requested by the HKMA.

5) Impact and risks

  • Operational: Potential for increased efficiency in alert generation and risk scoring, but requires ongoing validation of AI outputs.
  • Compliance: Need to ensure that AI-driven models remain explainable and auditable to satisfy supervisory expectations.
  • Data/IT: Shift towards holistic data integration necessary to support end-to-end AI transformation.

6) Compliance action checklist

  • [ ] Review the Annex report for specific industry benchmarks.
  • [ ] Conduct a gap analysis between current AI systems and the "holistic" models mentioned in the circular.
  • [ ] Update internal AI implementation roadmaps.
  • [ ] Brief relevant risk and compliance leads on the upcoming developments in Agentic AI.
  • [ ] Establish a process for archiving and updating progress notes for future HKMA reporting.

7) Appendices/attachments summary

  • Annex (Supporting Adoption of Artificial Intelligence in Fighting Financial Crime): This report provides a detailed analysis of AI deployment approaches across different sizes of AIs. It includes specific use cases (dynamic risk assessment, facial watch lists, corporate mule detection) derived from November 2025 workshops and follow-up technical sessions to guide AIs in their digital AML/CFT transformation.
中文摘要
快速切換摘要區塊
管理層摘要
  • 目的/背景 為持續推動人工智慧(AI)在反洗錢(AML)與反恐融資(CFT)領域的應用,金管局發布該報告,旨在總結認可機構(AIs)目前的 AI 部署策略、實務案例及成效,協助業界提升監控效率。
  • 一句話結論 金管局鼓勵各機構參考報告中的 AI 實務案例,持續優化現有的 AML/CFT 監控模型,並需按要求維持及更新內部 AI 實施計劃。
  • 關鍵變更
  1. 確立 AI 應用重點: 涵蓋動態風險評估、人臉識別監控名單及企業「傀儡戶口」(mule)檢測。
  2. 推動營運模式轉型: 鼓勵機構採納端到端(End-to-end)的轉型方案,將 AI 整合進整體洗錢監控架構。
  3. 強化重點監管方向: 未來關注焦點轉向「代理 AI」(Agentic AI)技術在金融犯罪偵測中的應用。
  • 重要日期 / 截止日
  • 2026 年 6 月 23 日: 舉辦「代理 AI」(Agentic AI)專題工作坊。
  • 即日起: 持續更新並審視 AI 實施計劃,並按金管局要求提交進度報告。
  • 適用對象 / 影響範圍 所有認可機構(AIs),特別是已提交 AI 實施計劃及有重大跨境/本地業務的機構。
  • 管理層建議行動
  1. 檢閱附件報告中的成功案例,評估將其應用於自身 AML 系統的可行性。
  2. 定期審視既有的 AI 部署實施計劃,確保與科技發展及合規要求同步。
  3. 積極參與金管局舉辦的相關技術研討會(如 2026/06/23 的 Agentic AI 工作坊)。
  4. 建立 AI 使用案例的內部存檔,以便隨時回應金管局的進度查詢。
詳細摘要
  1. 文檔概述
  • 性質: 監管通函及相關研究報告。
  • 目的: 總結業界 AI 應用現狀,提供技術支援與案例參考,推動 AML/CFT 系統升級。
  • 適用範圍: 全港所有認可機構(AIs)。
  1. 主要要求
  • 實務應用 機構應參考報告內關於動態風險評估、人臉識別及企業傀儡戶口檢測的技術框架。
  • 營運轉型 鼓勵機構超越單點應用,採取「整體性」(Holistic)方法,將 AI 整合進端到端的反洗錢營運模式。
  • 持續更新 已提交實施計劃的機構需保持計劃的時效性,並隨時準備向金管局報告新開發的使用案例(Use Cases)。
  1. 關鍵變更
  • 由基礎自動化轉向更複雜的「代理 AI」(Agentic AI)應用研究。
  • 強調從單一系統優化轉向全面性的反洗錢操作架構變革。
  1. 重要日期與過渡安排
  • 2026 年 6 月 23 日: 香港科技大學舉辦「代理 AI」實務工作坊。
  • 現行要求: 機構需根據過往提交的實施計劃執行,並按個別監管要求進行進度匯報。
  1. 對機構的影響與風險
  • 合規/營運 機構需持續評估 AI 模型準確性與可解釋性,避免模型偏移風險。
  • 數據/IT 涉及端到端模型轉型,需確保數據品質與處理效率能配合 AI 運作需求。
  1. 合規動作清單(Checklist)
  • [ ] 閱讀並分發附件報告至 AML 及 IT 技術團隊。
  • [ ] 評估機構內現有的 AML 模型,識別可運用 AI 優化之領域。
  • [ ] 審閱並更新內部的「AI 實施計劃」。
  • [ ] 準備相關進度報告,以備金管局隨時要求審查。
  • [ ] 指派代表參與 2026 年 6 月 23 日的工作坊。
  1. 附件/附錄摘要
  • 附件(Report on Supporting Adoption of AI in Fighting Financial Crime) 本報告匯總了 2025 年 11 月以來的工作坊成果,詳細分析了不同規模機構的 AI 部署路徑與實務應用案例,是機構推動技術轉型的關鍵指南。