Hong Kong’s financial and banking sector regulator has launched a generative artificial intelligence (AI) sandbox to foster collaboration between regulators, financial institutions and technology companies, with a focus on risk management, anti-fraud, customer experience and “AI vs. AI” strategies.
GenAI Sandbox++ was launched in early March by the Hong Kong Monetary Authority (HKMA), Securities and Futures Commission (SFC), Insurance Authority (IA), and Mandatory Provident Fund Authority (MPFA) in collaboration with Hong Kong Cyberport Management Company (Cyberport), the city’s digital technology “flagship and AI accelerator.”
The scheme builds on the initial GenAI Sandbox, launched in 2024, which provided a risk-controlled environment for authorized institutions in the banking sector to develop and test innovative AI-based solutions. We also considered AI vs. AI strategies to manage risks related to AI deployments, including the use of AI to perform automated quality checks on the output generated by AI.
According to the SFC, the “strong industry response to the first cohort” of the 2024 Sandbox demonstrated “the significant potential of AI technologies in improving operational efficiency, risk management, and customer engagement.”
The newly announced Sandbox++ will therefore expand the project to cover multiple financial sectors including banking, securities and capital markets, asset and wealth management, insurance, mandatory provident funds (MPFs), and stored value facilities, while continuing to drive AI-to-AI strategies to “manage risks associated with AI deployments”.
The HKMA said participating financial institutions will receive targeted supervisory guidance, technical support and free access to graphics processing unit (GPU) computing resources at Cyberport’s AI Supercomputing Center. The aim is to enable participants to develop, pilot and refine use cases “in a risk-managed environment”, thereby accelerating responsible AI adoption across Hong Kong’s financial ecosystem.
By fostering collaboration between regulators, financial institutions and technology companies, the group behind the initiative hopes to “generate new ideas and foster deeper partnerships across sectors and borders.”
For example, we will encourage the development of sector-specific and cross-sector AI applications, such as AI-powered underwriting and claims processing, AI-powered tools for processing MPF (pension funds), and industry-wide use cases such as intelligent customer chatbots and advanced fraud detection systems.
“The launch of GenAI Sandbox++ marks a key milestone in our ‘FinTech 2030’ strategy and reinforces our commitment to building a vibrant ecosystem for responsible innovation,” said Eddie Yue, Chief Executive of HKMA. “By bringing together regulators, financial institutions and the technology community, we aim to unlock the full potential of AI, drive growth, efficiency and customer-centricity across financial services, and further strengthen Hong Kong’s competitiveness as a leading international financial centre.”
SFC CEO Julia Leon echoed this sentiment, saying, “Expanding the GenA.I. Sandbox to a wider range of capital market participants is a true testament to our collective determination to drive future-proof financial markets powered by responsible market innovation and technology.”
Applications for the sandbox are being accepted until June 30, 2026, and Leong encourages licensed companies to “actively participate in this sandbox to leverage the limitless potential of AI in improving operational efficiency, strengthening resilience, and growth potential.”
For artificial intelligence (AI) to function properly within the law and succeed in the face of growing challenges, it must integrate enterprise blockchain systems that ensure the quality and ownership of data input. This allows you to keep your data safe while ensuring data immutability. To learn more about why enterprise blockchain is the backbone of AI, check out CoinGeek’s coverage of this emerging technology.
See: Solutions to overcome blockchain regulatory hurdles
