-
The emergence of generative artificial intelligence ("GenA.I."), powered by large language models ("LLMs") that facilitate seamless human-A.I. interaction, marks a pivotal advancement in the evolution of artificial intelligence ("A.I.") technology. At its core, GenA.I. possesses the capabilities to generate human-like, novel outputs across a variety of media such as text, images, audio and video with profound implications for numerous industries. Among these, the financial services sector stands to be particularly transformed by GenA.I., given its potential to revolutionise operational efficiency, risk management and customer engagement.
GenA.I.'s rapid growth is fuelled by advances in computational power and increasingly sophisticated model architectures, enabling financial institutions to streamline key functions. Two key areas are (1) knowledge management and (2) creative content management processes, which are enhancing operational efficiency.
To gauge the current adoption efforts of GenA.I., the HKMA conducted an industry-wide survey participated by 137 industry practitioners and in-depth interviews with 16 organisations across the technology, banking, securities, and insurance sectors. Findings reveal that most financial institutions are currently focusing on internal applications that drive employee productivity. These include innovations aimed at automating routine tasks, improving workflow efficiencies, and enhancing knowledge sharing. The adoption of GenA.I. for customer-facing applications remains nascent. Concerns around regulatory compliance, operational deployment, and the resource-intensive nature of GenA.I. solutions are cited as key barriers to more widespread adoption.
In response to rapid advancements in GenA.I., regulators globally have been revisiting or updating their frameworks to ensure that the deployment of A.I. technologies is both responsible and ethical. Some jurisdictions, including the European Union (EU), Hong Kong Special Administrative Region (SAR), Mainland China, Singapore, United Kingdom (U.K.), and United States (US) are refining regulations or supervisory guidance to strike a balance between innovation and consumer protection. These frameworks emphasise critical principles such as: - Governance & Accountability
- Fairness
- Data Privacy & Protection
- Transparency & Disclosure
-
While the potential of GenA.I. to transform the financial services sector is immense, its adoption also introduces several key risks that financial institutions must address to ensure safe and responsible deployment. These include concerns around inaccuracy or hallucination, where A.I. models may generate incorrect or misleading information, as well as data privacy risks due to the large datasets often required by GenA.I. systems. Additionally, risks related to algorithmic bias, cybersecurity vulnerabilities, and operational challenges such as system integration and scalability must be carefully managed to avoid negative impacts on both operations and reputation.
Against this regulatory backdrop, financial institutions are encouraged to establish a robust GenA.I. governance structure that ensures compliance and facilitates the responsible deployment of new solutions. This governance framework should include the development of business cases and proof-of-concepts that are complemented by comprehensive strategies for risk management, system integration, and continuous monitoring. A well-designed feedback loop across all phases of adoption is essential to drive ongoing improvements and maximise the value of GenA.I. initiatives while mitigating risks.
Recognising the resource intensity required to implement A.I and GenA.I. solutions, Hong Kong has introduced a range of facilitation initiatives to support financial institutions. These includes access to high-performance computing resources, financial subsidies, talent development programmes, and promotional events aimed at fostering innovation and building a skilled workforce capable of leveraging GenA.I. technologies.
