As artificial intelligence (“AI”) technologies continue to permeate the financial sector, regulators worldwide have faced challenges in addressing the potential risks associated with their adoption.
The Hong Kong Monetary Authority (“HKMA”) has taken a multifaceted approach to regulating AI in Hong Kong’s financial sector. Its approach to AI regulation is characterized by a principle-based, rather than prescriptive, framework that balances fostering innovation with mitigating risks. The HKMA’s strategy has evolved over time, adapting to technological advancements and international best practices.
Early Stages: High-Level Principles
In 2019, the HKMA issued a circular outlining high-level principles for authorized institutions (“AIs”) using AI applications. The 12 principles were designed to be flexible rather than prescriptive, allowing for proportionate application based on AI usage and risk levels. The principles are grouped into three main categories:
- Governance
- Board and senior management accountable for the outcome of AI applications
- Application Design and Development
- Possessing sufficient expertise
- Ensuring an appropriate level of explainability of AI applications
- Using data of good quality
- Conducting rigorous model validation
- Ensuring auditability of AI applications
- Implementing effective management oversight of third-party vendors
- Being ethical, fair and transparent
- Ongoing Monitoring and Maintenance
- Conducting periodic reviews and on-going monitoring
- Complying with data protection requirements
- Implementing effective cybersecurity measures
- Risk mitigation and contingency plan
On 23 December 2019, the HKMA released a report on AI use in Hong Kong’s banking sector, featuring insights from academics and industry experts. It revealed that nearly 90% of surveyed retail banks had adopted or planned to adopt AI, with 95% of those banks intending to use AI to shape their corporate strategy, driven by goals to enhance customer experience, maintain cost-effectiveness, and improve risk management.
Expanding Focus: Suspicious Activity Monitoring and Broader Support
By 2024, the HKMA’s focus broadened. By a circular dated 9 September 2024, the HKMA encouraged AIs to utilize AI for enhancing the monitoring of money laundering and terrorist financing (“ML/TF”) risks. Recognizing the increasingly complex ML/TF landscape, the HKMA highlighted the advantages of AI-powered systems over conventional rules-based systems.
This reflects a shift towards leveraging AI’s capabilities to enhance the effectiveness and efficiency of AML/CFT compliance. The HKMA supports AI adoption through experience-sharing forums, dedicated supervisory teams providing technical guidance, and fostering a conducive environment for innovation within the Fintech Supervisory Sandbox and Chatroom. AIs are required to conduct feasibility studies and submit implementation plans by March 2025.
On 13 August 2024, the HKMA launched a new Generative Artificial Intelligence (“GenAI”) sandbox in collaboration with Cyberport, allowing banks to test new AI ideas within a risk-managed framework. This reflects HKMA’s iterative, step-by-step approach in promoting the application of GenAI in the banking industry.
During Hong Kong FinTech Week 2024, the Chief Executive of the HKMA Mr Eddie Yue emphasized the importance of ongoing innovation in reshaping the fintech landscape and called for the financial sector to remain adaptable and open to new opportunities. He reiterated HKMA’s commitment to unlocking fintech’s full potential to drive sustainable growth within the ecosystem.
HKMA also plans to introduce a sector-wide uplift programme to enhance banks’ monitoring of complex money laundering cases and geopolitical risks through the use of AI, supported by a joint event with Cyberport.
Parallel Government Initiatives
The government’s actions, as seen in the “Policy Statement on Responsible Application of Artificial Intelligence in the Financial Market” (the “Policy Statement”) issued by the Financial Services and Treasury Bureau (“FSTB”) on 28 October 2024, demonstrated a parallel commitment to responsible AI development. This includes addressing potential risks like cybersecurity, data privacy, and intellectual property protection, while promoting international collaboration and supporting the growth of the fintech sector.
The Policy Statement acknowledged Hong Kong’s high AI adoption rate in financial institutions (38% compared to the global average of 26%) and introduced a dual-track approach that aims to both promote AI development and address potential risks. The policy recognized AI’s key applications in areas such as research, investment management, customer service, risk assessment, fraud prevention, and workflow automation.
The Policy Statement outlined comprehensive risk mitigation strategies focusing on cybersecurity, data privacy, model transparency, and financial stability. To support responsible AI adoption, the government plans to develop clear supervisory frameworks, provide access to AI models like HKUST’s InvestLM, and collaborate with financial regulators.
In summary, the HKMA’s approach to AI regulation is a dynamic one, characterized by a principle-based framework that evolves with technological advancements and industry needs. It balances the promotion of innovation with risk mitigation, actively supporting responsible AI adoption while addressing potential challenges through guidance, support, and collaboration with the government and industry stakeholders.