AI Tracker
Hong Kong

Financial Services

There is no central framework for regulating the use of AI in the financial services sector in Hong Kong. Instead, the Securities and Futures Commission (SFC), Financial Services and the Treasury Bureau (FSTB), and the Hong Kong Monetary Authority (HKMA) have published circulars and guidelines to outline their regulatory expectations in this area.

• The FSTB issued a Policy Statement on the Responsible Application of Artificial Intelligence in the Financial Market in October 2024, setting out the government's 'dual-track' approach towards the responsible adoption of AI in the financial market and addressing potential challenges. The policy statement provides that financial institutions should formulate an AI governance strategy and adopt a risk-based approach in the procurement, use, and management of AI systems, with human oversight to mitigate the potential risks.
• In November 2024, the SFC issued a circular on the use of generative AI language models that set out its expectations for licensed corporations that offer services or functionality provided by Generative AI language models (including as part of third party products) in respect of their regulated activities. It identified four core principles that licensed corporations may implement in a risk-based manner, specifically senior management responsibilities, AI model risk management, cybersecurity and data risk management, and third party provider risk management; the Appendix to the circular further lists non-exhaustive risk factors tied to each core principle. Aside from the circular, the SFC is participating in the International Organisation of Securities Commissions' Fintech Task Force AI Working Group - as part of which it will keep under review any findings or recommendations from such organisation to consider whether further regulatory guidance to SFC-licensed firms is necessary.
• In November 2019, HKMA issued a circular that set out high-level principles to the banking industry on the use of AI applications, covering their governance, application design and development, and monitoring and maintenance. These principles are supplemented by publications including:
  • A report titled Fintech Adoption: Progress and Future Directions published on 16 July 2025, which highlights key observations from a tech maturity stock-take exercise (including significant advancements in the adoption of AI by the banking industry), and indicates plans to formulate a detailed blueprint of initiatives to further support the industry's adoption of sophisticated fintech, including AI.
  • A research paper titled Generative Artificial Intelligence in the Financial Services Space, published on 27 September 2024, which explores the transformative potential of generative AI and its implications for the financial industry, particularly in terms of operational efficiency, risk management, and customer engagement.
  • A circular and Annex, issued on 9 September 2024, to explain how the use of AI can improve the effectiveness and efficiency of monitoring money laundering and terrorist financing (ML/TF) risks. It also sets out the actions that the HKMA is taking to further support and accelerate the use of AI in authorised institutions’ monitoring processes, for example, establishing a dedicated team (supported by an external consultant) to provide supervisory feedback and technical guidance to assist authorised institutions in applying AI in enhancing their monitoring processes. The circular also requested authorised institutions with significant operations in Hong Kong to give due consideration to adopting AI in their ML/TF monitoring systems, and submit a feasibility study report and AI implementation plan to the HKMA by end of March 2025.  In a circular dated 19 November 2025, the HKMA noted that 48 authorised institutions had conducted the feasibility study and the vast majority agreed that AI was useful in making ML/TF monitoring systems more effective and risk-based.  To support authorised institutions' AI implementation efforts the HKMA has commenced a series of workshops on topics including approaches to innovating monitoring systems, AI use cases, and crypto asset-related ML/TF risks.  The HKMA has also observed in its 16 March 2026 circular the use of AI by authorised institutions to optimise sanctions screening processes.
  • A circular, issued on 19 August 2024, to provide guiding principles on the use of GenAI in customer-facing applications from a consumer protection perspective. These build on the guiding principles the HKMA issued on 5 November 2019 in respect of the use of big data analytics and AI, focussing on four key aspects: governance and accountability, fairness, transparency and disclosure, and data privacy and protection. 
  • An insight article, shared on 23 May 2024, regarding the implications of AI on manpower management in the banking industry. In this context, the HKMA will publish an updated study on "Capacity Building for Future Banking" in 2025 to identify skills gaps in the banking industry for the next five years (2026 - 2030).
  • A thematic review, published on 17 April 2024, of the design, implementation and optimisation of transaction monitoring systems. The review included an examination of how AI is used to optimise transaction monitoring systems and provided anti-money laundering and counter-financing of terrorism specific guidance based on industry best practices.

From a more practical perspective, the HKMA has invited authorised institutions to participate in the GenAI Sandbox initiative launched in August 2024 in collaboration with the Hong Kong Cyberport Management Company Limited (Cyberport). The GenAI Sandbox aims to promote responsible innovation in GenAI and empowers banks to pilot GenAI use cases within a risk-controlled environment across three critical domains, risk management, anti-fraud measures, and customer experience. The first cohort of the GenAI sandbox was announced in December 2024 and a report was published on 31 October 2025 highlighting the key findings and providing guidance to help banks address key challenges throughout the AI implementation lifecycle, including data preparation, model fine-tuning, output evaluation, and ongoing monitoring and optimisation.  The second cohort of the GenAI sandbox was announced on 15 October 2025, with use cases on "AI vs AI" strategies and advanced defence mechanisms against deepfake-related fraud.  On 5 March 2026, various financial regulators, including the SFC, the HKMA, the Insurance Authority, and the Mandatory Provident Fund Schemes Authority announced the launch of the GenAI Sandbox++, an expanded version of the 2024 sandbox covering multiple financial sectors under the remit of such regulators. 

Separately, the HKMA indicated in September 2025 that it was collaborating with a few like-minded central banks and regulators on a joint research project on an explainable AI toolkit, Project Noor, which was initiated by the Bank of International Settlements.

During the FinTech Week in November 2025, the HKMA announced "Fintech 2030", a forward-looking strategy for driving Kong Kong's fintech development, of which one of the four strategic pillars will focus on a new holistic AI strategy for authorised institutions.  As part of the "Fintech 2030" strategy, the HKMA published a Fintech Promotion Blueprint in February 2026, setting out a tactical framework to foster responsible innovation and advance fintech development in the banking sector.  One of the five key technology enablers and foundations in the blueprint is AI.

On 9 March 2026, the HKMA issued a circular requiring boards of directors of authorised institutions to oversee and endorse a formal strategic business plan, by 9 September 2026, to identify opportunities where technology can be used to adapt or transform product offerings, revenue models, customer engagement, risk management and operations.  A selection of authorised institutions will be invited to submit their plans via HKMA's survey tool.

The HKMA itself also utilises tech tools for conduct supervision, such as a workflow system powered by agentic AI and web-scraping technologies to detect and monitor misleading marketing claims, including the misuse of terms like 'bank' or 'deposit.

In addition:

• On 3 November 2025, the Insurance Authority announced the publication of a whitepaper on federated learning, a machine learning technique that enables AI models to be trained on decentralised datasets residing on devices or servers, thereby complying with data privacy regulations while facilitating collaboration across different entities and sectors.
• On 18 August 2025, the Insurance Authority indicated that updated guidelines will be issued in 2026 to provide regulatory clarity on the use of AI and facilitate AI adoption in a responsible way without creating unnecessary regulatory frictions or hurdles. It also launched the AI Cohort Programme to support insurers at all stages of technological maturity and to accelerate the responsible use of AI across the insurance industry. Seven insurers pledged as core participants of the AI Cohort and are committed to establishing an AI Centre of Excellence in Hong Kong to serve as a regional hub for AI innovation, supporting AI talent development through capability-building initiatives and contributing to knowledge sharing with the industry and regulators.
• In 2023, the Insurance Authority outlined the responsibilities of insurers and intermediaries regarding the use of chatbots under the 'regulated activities' regime, reminding them to uphold the principles of fair customer treatment and enable fully informed customer decisions. The Insurance Authority is enhancing its Guideline on Cybersecurity by developing a Cyber Resilience Assessment Framework specific for the insurance sector. It will also conduct a Fintech survey to evaluate current technology adoption trends in the insurance sector, focusing on AI and related cybersecurity measures.

Healthcare

On 3 January 2024, Hong Kong’s Medical Device Division (MDD) of the Department of Health issued a Technical Reference document TR-008: Artificial Intelligence Medical Devices which provides clarity for devices using AI and machine learning - including continuous learning capability - and the technical requirements expected for listing these medical devices on the Medical Device Administrative Control System.  This document applies to all AI medical devices that fall within the scope of the Medical Device Administrative Control System.