tech · HK01

HKMA Unveils ART Framework to Guide Responsible AI Adoption in Banking

1 day ago2 MIN
HKMA Unveils ART Framework to Guide Responsible AI Adoption in Banking

Summary

The Hong Kong Monetary Authority (HKMA) unveiled the "ART" framework on October 15, 2026, marking a significant step toward establishing clear regulatory guidance for artificial intelligence adoption in the banking sector. The framework, designed to foster an accountable, resilient, and trustworthy AI-enabled banking ecosystem, will replace the authority's 2019 High-Level Principles on AI .

Key Points

  • The ART framework encompasses three core pillars: Accountability, Resilience, and Trust, providing banks with a structured approach to AI governance
  • The HKMA will release a consultation paper on Agentic AI governance in the coming weeks, covering accountability arrangements, risk management, and operational scope
  • Second-phase Generative AI sandbox results showed 27 use cases from 20 banks and 14 technology partners, with positive outcomes demonstrating practical operational benefits
  • Specific implementations include risk analysis AI agents reducing query time by 95% with 100% accuracy, and eKYC systems cutting account opening time by 40%
  • The HKMA is collaborating with Cyberport and the University of Hong Kong to develop an AI training data sharing platform with data quality assurance as a core design principle

Why It Matters

As AI systems become more autonomous and capable of mobilizing greater data, tools, and workflows, the ART framework provides Hong Kong banks with a clear regulatory foundation to innovate responsibly while maintaining financial system stability and public trust in the sector .
As AI systems become more autonomous and capable of mobilizing greater data, tools, and workflows, the ART framework provides Hong Kong banks with a clear regulatory foundation to innovate responsibly while maintaining financial system stability and public trust in the sector .

READ IT IN THE APP

Download on the App Store