Supporting Artificial Intelligence Adoption in AML/CFT

55th Floor, Two International Finance Centre, 香 港 中 環 金 融 街 8 號 國 際 金 融 中 心 2 期 55 樓 8 Finance Street, Central, Hong Kong 網 址：www.hkma.gov.hk Website: www.hkma.gov.hk Our Ref.: B10/1C B1/15C 19 November 2025 The Chief Executive All Authorized Institutions Dear Sir/Madam, Supporting Artificial Intelligence Adoption in AML/CFT Further to our letter of 9 September 2024 regarding the use of artificial intelligence to monitor suspicious activities, I am writing to provide an update on the support the Hong Kong Monetary Authority (HKMA) will provide in this area and other aspects of anti￾money laundering and counter financing of terrorism (AML/CFT). Forty-eight Authorized Institutions (AIs) have, in response to our letter, assessed the feasibility of deploying artificial intelligence to strengthen the effectiveness of transaction monitoring. The vast majority agree with the HKMA on the usefulness of artificial intelligence to make money laundering and terrorist financing (ML/TF) monitoring systems more effective and risk-based, thereby enhancing the sector’s collective resilience to the evolving threat landscape through intelligence-focused preventive measures. More than 30% of AIs have already adopted artificial intelligence as part of their monitoring systems, while most others have provided timelines along which they will do so and which will result in the adoption rate rising to above 80% over the course of the next 12-24 months. Different approaches were noted in how artificial intelligence is being adopted. Some AIs have replaced reliance on rules-based transaction monitoring systems by transitioning to a more holistic approach. This involves transforming the overall data structure, making better use of technology, and applying more advanced analytics to a wider range of customer and transactional data to identify and disrupt threats. Other AIs, while sharing the same overall objective, are implementing artificial intelligence as part of a phased approach, adopting certain use cases as building blocks to make existing processes more effective and efficient. Use cases are largely concentrated in risk detection and transaction monitoring alert prioritisation, suggesting that there is scope for broader and deeper adoption. Some AIs also participated in the Generative Artificial Intelligence Sandbox to explore AML and anti-fraud use cases.

• 2 -
  To support these efforts, the HKMA plans to roll out a series of workshops covering analysis of the different approaches to innovating monitoring systems, as well as use cases involving artificial intelligence in risk detection, alerts prioritisation and the use of generative artificial intelligence to compile suspicious transaction reports. Considering an increasing number of AIs are planning to provide crypto asset-related services, an additional workshop focused on ML/TF risk management in this area will also be organised. These workshops will be accompanied by follow-up support and publications providing advice and recommendations. The first of this series of workshops analysing the relative merits of adopting a holistic approach vis-à-vis a building block approach to innovating systems was held today at Cyberport. The HKMA will continue to track progress of adoption according to implementation plans submitted and keep the effectiveness of banks’ transaction monitoring systems under regular review. AIs may approach their usual supervisory contact in the HKMA’s AML and Financial Crime Risk Division on any questions relating to this circular. Yours faithfully, Raymond Chan Executive Director (Enforcement and AML)