The Hong Kong Monetary Authority has launched the Accountability, Resilience and Trust Framework to support responsible artificial intelligence adoption in banking as systems become more autonomous and gain greater access to data, tools and workflows. The framework advances the responsible AI agenda under the authority’s Fintech 2030 strategy, combining governance and operational resilience measures with shared infrastructure and further testing of agentic AI. The package includes updated supervisory guidance informed by the second cohort of the Generative AI Sandbox and a training data sharing platform developed with Hong Kong Cyberport Management Company Limited and the University of Hong Kong, with data quality assurance as a core design principle. The authority will also test agentic AI applications through the Generative AI Sandbox++ and publish a discussion paper on a proposed governance framework covering intent anchored accountability, a registry of material AI agents and effective intervention mechanisms. More than 60% of use cases in the completed second sandbox cohort involved agentic AI, including automated planning and execution of multistep tasks within defined boundaries and controls.
Hong Kong Monetary Authority launches ART Framework for responsible artificial intelligence adoption in banking
The Hong Kong Monetary Authority has launched the Accountability, Resilience and Trust Framework for responsible AI adoption in banking under its Fintech 2030 strategy. It combines updated supervisory guidance and a training data sharing platform with further agentic AI testing. An upcoming discussion paper will explore accountability, a registry of material AI agents and intervention mechanisms.