The Hong Kong Securities and Futures Commission (SFC) has published its first AI adoption survey, setting out four key regulatory priorities

Main text

On November 2, the Securities and Futures Commission of Hong Kong (the "SFC") published its first "Survey on the Use and Preparedness of Artificial Intelligence by Large Institutions" (the "Survey"). The results show that institutions have widely applied artificial intelligence (AI) across business and monitoring functions, but scaling up its use responsibly still faces challenges.

In the Survey, the SFC also disclosed that it will adopt a regulatory approach that is safety-first, people-oriented and collaboration-led, focusing on addressing emerging security, cybersecurity and resilience risks; maintaining human accountability while granting AI more authority; and promoting collaboration with other regulators, the industry, technology vendors and academia to fill gaps in infrastructure, knowledge and capability.

95% of surveyed institutions expect to accelerate AI deployment

According to the introduction, the survey was conducted between July and August 2026, covering 20 large financial institutions and their 78 affiliated licensed corporations in Hong Kong. The responses show that AI use is permeating business and monitoring functions, integration with internal systems is increasing, and some applications have become more autonomous.

The survey results particularly note that agentic AI (AI capable of autonomously deciding and executing steps to achieve a goal, including accessing data or tools and, once authorized, taking action under limited human instruction) is becoming increasingly common, mainly used for information retrieval, content preparation, software development, operational workflows and some trading-related activities. 30% of surveyed institutions have already applied agentic AI in business operations, while another 25% are conducting pilots, involving varying degrees of autonomy and human involvement.

Other key survey findings include:

1. AI use is gaining momentum, with 95% of surveyed institutions expecting to accelerate AI deployment in the next 12 months. Its application now covers front-, middle- and back-office activities, and the degree of system integration continues to rise, with 90% of institutions having connected AI tools to internal systems.

2. Institutions are using AI for cyber defense, with 90% adopting frontier AI models for this purpose or evaluating such applications, accelerating analysis and expanding monitoring coverage.

3. The practical difficulties facing wider AI adoption include technical integration, moving pilot projects into formal operation, uncertainty over costs and business value, data limitations, and skills shortages. As applications continue to increase, institutions face various challenges, including the need to expand the scope of governance and monitoring measures while maintaining clear accountability arrangements and human oversight.

The SFC said the survey results will help it formulate its AI regulatory approach, which has four key priorities: (1) protecting investor interests as AI reshapes how investors interact with institutions; (2) strengthening cybersecurity and operational resilience to address new threats and dependencies brought by AI; (3) re-examining how to implement human accountability and human oversight in increasingly autonomous workflows to ensure effective AI governance; and (4) building knowledge and capability through exchange and collaboration.

Ye Zhiheng, Executive Director of the SFC's Intermediaries Division, said: "We adopt a regulatory approach that is safety-first, people-oriented and collaboration-led. As AI becomes increasingly autonomous, we must ensure that human primacy is not undermined. To promote scaled AI adoption, strong models and promising use cases alone are not enough; reliable infrastructure, robust data, appropriate professional expertise, and a governance system that keeps pace with technological development are also needed."

At an event on October 30, SFC Chairman Huang Tianyou said that financial markets have never been built on technology alone, but on trust, and behind trust are clear responsibilities and accountability. What the financial industry truly needs is not just AI, but Accountable Intelligence. AI can help people make judgments, but it cannot replace people in bearing responsibility; functions can be delegated, but responsibility cannot be outsourced.

In Huang Tianyou's view, truly achieving "Accountable Intelligence" has three key points.

First, it must be "visible." Boards and senior management must first clearly know where institutions use AI, what judgments AI is making, and what risks it may bring. AI cannot be merely an information technology (IT) department project; it needs to be incorporated into the institution's overall risk management and governance framework.

Second, it must be "manageable." The stronger AI's capabilities, the more important human oversight becomes. Especially for major decisions involving client interests, investment decisions, risk-taking or compliance responsibilities, we must retain genuine and effective human oversight. Management cannot hand responsibility to machines simply because "AI made the decision."

Third, it must be "questionable." Not every one of our employees needs to become an AI expert, but at least they must be able to understand how AI affects their work, be able to question AI outputs, and know when they cannot rely on AI alone. Equally important, management must be able to explain and account to boards, clients and even regulators: why use AI? How are risks managed? If problems arise, who is responsible?

Huang Tianyou particularly emphasized: "Good AI governance and innovation are not opposites. On the contrary, the clearer the governance and accountability, the more conditions we have for bold innovation. Good governance allows innovation to proceed more steadily and go further. Only when responsibilities are clear and risks are controllable will clients have confidence, the market accept it, and technology investment have the chance to truly translate into corporate competitiveness. For Hong Kong to build an internationally leading fintech ecosystem, we must certainly be bold in innovation, but we must also continue to protect the most important asset of Hong Kong's financial market, and that is trust."

In addition, the SFC reminded industry institutions to comply with applicable laws and regulatory requirements to support the responsible use of AI in Hong Kong's capital markets. The SFC will focus on ensuring that human accountability is maintained throughout business processes using AI, including setting clear objectives, clearly defining the scope of data access and authorized actions, and establishing effective monitoring and intervention mechanisms. Relevant control measures should be commensurate with the authority and potential impact of the application, rather than requiring every output or action to be approved by a human.

(Source: Securities Times) Middle of the main text Embedded advertisement