HKMA: Focused on how banks effectively govern AI, will issue guidelines on practical applications.

Speaking at the Hong Kong Bankers Summit, HKMA Deputy Chief Executive Arthur Yuen said artificial intelligence is no longer merely assistive in decision-making and action it can execute. He is focused on how banks effectively govern AI, how to enhance technological resilience and ensure the banking system can evolve in step with an increasingly complex environment, how to build a trustworthy tokenisation framework, and how to ensure data quality. He also said the HKMA will issue guidelines on practical applications to help banks translate data excellence principles into concrete actions. Arthur Yuen noted that AI can provide decision options and execute parts of processes, but this does not mean every application scenario must be automated or autonomous. He stressed the importance of human oversight, pointing out that ultimately humans must still make decisions and bear responsibility. Tokenisation makes the transfer of value programmable, changing how transactions are structured and how financial services are delivered to users, creating substantial opportunities for banks. But he stressed that trust must not be sacrificed as a cost, and it must be ensured that all transactions conducted through blockchain are genuine, accurate and authorised. He emphasised that the HKMA still believes frontier AI models have not fundamentally changed the landscape of cyber risk management, but standards must be adjusted to address faster and larger-scale cyber attacks. The authority is actively guiding the industry to address these risks, and also expects banks to strengthen cybersecurity controls and assess whether sufficient risk management measures are taken for services provided by third parties, and whether this could create backdoor attack risks. He mentioned that the HKMA launched a generative AI sandbox several years ago, testing more than 40 use cases, half of which have already been productised.