Generative Artificial Intelligence Sandbox: Second Cohort

55th Floor, Two International Finance Centre, 香 港 中 環 金 融 街 8 號 國 際 金 融 中 心 2 期 55 樓 8 Finance Street, Central, Hong Kong 網 址：www.hkma.gov.hk Website: www.hkma.gov.hk Our Ref.: B1/15C 28 April 2025 The Chief Executive All Authorized Institutions Dear Sir/Madam, Generative Artificial Intelligence Sandbox: Second Cohort I am writing to invite Authorized Institutions (AIs) to apply for the second cohort of the Generative Artificial Intelligence (GenA.I.) Sandbox initiative launched by the Hong Kong Monetary Authority (HKMA) in collaboration with the Hong Kong Cyberport Management Company Limited (Cyberport) 1 . The GenA.I. Sandbox is designed to provide a risk-controlled environment for AIs to develop and test innovative artificial intelligence-based solutions. The number of proposals received for the first cohort 2 reflected the strong interest within the banking sector in harnessing A.I. technology for innovation. As the first cohort progresses, with its expected conclusion later this year, the HKMA is pleased to announce the launch of the second cohort in response to the sustained interest. Specifically, the GenA.I. Sandbox Collaboratory is introduced to further foster collaboration between AIs and fintech firms in developing A.I.-related use cases. The Collaboratory features a series of practical workshops designed to explore targeted use cases in the Banking sector and how A.I. solutions can effectively address AIs’ problem statements through ideation and prototyping. The outputs of the Collaboratory are expected to assist AIs in formulating their use cases for the second cohort. /…page 2 1 HKMA Press Release, “HKMA and Cyberport Launch GenA.I. Sandbox to Bolster A.I. Adoption in Financial Sector” issued on 13 August 2024 2 HKMA Press Release, “HKMA announces inaugural cohort of GenA.I. Sandbox” issued on 19 December 2024

• 2 -
  Given the growing importance of managing risks associated with A.I. adoption, all use cases in the second cohort are expected to incorporate robust A.I. safety validation into their technical trials and evaluations. The development of use cases with a component of A.I.-assisted second and third lines of defence to enhance A.I. governance is encouraged. AIs may refer to the updated catalogue of Fintech firms shared by the Cyberport for solutions focusing on risk management of A.I. when forming partnerships for their second cohort use cases. Applications for the second cohort of the GenA.I. Sandbox are now open and will be accepted until 31 August 2025. Similar to the first cohort, applications will be prioritised based on a number of factors (please refer to the Annex for details). Selected projects will tentatively be announced in the fourth quarter of this year. Should you have any questions regarding the Sandbox or application process, please contact Mr. John Chiu or Ms. Joyce Ip of Banking Supervision Department at GenAI_sandbox@hkma.gov.hk. Yours faithfully, Carmen Chu Executive Director (Banking Supervision) Encl

---

Annex

• 1 -
  Arrangement for the Second Cohort of Generative A.I. Sandbox (“Sandbox”) Objectives

• The Sandbox is intended to:

• support the development, testing, and pilot of innovative Artificial
  Intelligence (A.I.) and Generative A.I. (GenA.I.)-based solutions in the banking sector; and

• provide early targeted supervisory feedback to participants in the Sandbox
  and draw insights from sandbox trials. Where appropriate, the Hong Kong Monetary Authority (HKMA) will further provide guidance and share good practices on the adoption of A.I. and GenA.I..

• The Sandbox use cases are expected to focus on enhancing risk management,
  anti-fraud measures and customer experience. Some non-exhaustive example use cases are shared below for inspiring innovative ideas:

• Risk management – Optimise risk management processes for the full
  range of risk disciplines relevant to AIs, such as to evaluate creditworthiness and on-going performance of borrowers by analysing financial statements and other unstructured data to enhance the efficiency of credit lifecycle; review business processes to improve operational resilience; and generate risk assessment reports and timely risk warning;

• Anti-fraud measures – Detect and prevent “Deepfake” scams, fraudulent
  document; review website, emails and message contents to automatically identify fraudulent messages; devise timely responses; and formulate preliminary investigation reports based on identified fraud patterns; and

• Customer experience – Elevate customer engagement through the use of
  more advanced customer service chatbots capable of generating personalised responses based on individual customer background, transaction records and past interactions; interact with customers through mediums other than text by leveraging multimodal capabilities of GenA.I.; and generate more timely and customised communications based on customer context and relevant external news.

• All Sandbox use cases should incorporate a component of A.I. risk
  management and safety. Examples may include bias detection and mitigation, explainable A.I., as well as frameworks for A.I. output monitoring and evaluation.

• 2 -
  General Principles

• The focus of the Sandbox is on solutions that demonstrate a significant level
  of innovation and potential for substantial impact. Priority will be given to solutions that align with these criteria.

• Adhering to the general principle of data minimisation, AIs are strongly
  encouraged to adopt data sensitisation techniques such as data masking or tokenisation on their training data, where applicable, to minimise the risk of data leakage, while still ensuring the models are trained and tested under realistic conditions.

• A secure data transfer and access mechanism will be made available to ensure
  the confidentiality and integrity of the data used within the Sandbox. AIs are expected to review and implement adequate data security controls, including encryption and access controls where necessary.

• The level of A.I. or GenA.I.-specific risk mitigations associated with the
  proposed solutions will be a key factor in the project selection. Projects that focus on exploring and testing a wide range of GenA.I.-driven risk mitigation strategies will be prioritised over simpler or commonly adopted cases, the latter of which may offer less learning or impact.

• All use cases are recommended to incorporate a robust A.I. safety validation
  component into their technical trials and evaluations.
  Key Factors for Consideration in Evaluating Applications

• Level of innovation – The extent to which the proposed solution introduces
  novel or alternative ideas, methodologies or models that have the potential to create new digitalisation opportunities;

• Complexity of the solutions – The technical sophistication and intricacy of
  the proposed solutions to promote advancement and added value;

• Expected contribution to the industry – The potential for the proposed
  solution to make a meaningful impact on the financial services sector, such as by addressing industry-wide challenges or by formulating new solutions replicable and scalable by different institutions; and

• Adherence to the principle of fair use – The degree to which the proposed
  solution is designed to be used in a fair, responsible, and ethical manner, avoiding harm to others or misuse of computing power.

• 3 -
  Application Procedure

• Interested parties can submit their applications through the HKMA’s Survey
  Tool platform. Applicants are required to provide details including high-level design, applicable models, preliminary risk assessments, approach and timeline of proposals and information about their technology partners.

• The HKMA may contact the applicant to gather additional information during
  the evaluation process. Processing time will depend on the complexity of the proposal, quality of the information provided, and responsiveness to follow￾up questions. Sandbox Collaboratory Arrangement

• The HKMA will issue separate invitations to AIs. Interested AIs may submit
  their areas of interest and potential use cases.

• AIs and technology vendors will be grouped into sessions based on their
  interests, areas of expertise and availability.

• AIs may identify their technology partners during these sessions to participate
  in the GenA.I. Sandbox for further exploration and more comprehensive testing following initial ideations. Others

• Admission to the sandbox does not indicate HKMA’s endorsement of the
  solution.

• Sandbox participants intending to implement the solution after the Sandbox
  should adhere to established procedures when adopting new technologies.

• The HKMA may publicly announce information regarding sandbox
  participation, including the disclosure of approved participants.

• The HKMA may adjust sandbox requirements based on participants’ proposals
  and testing progress.
  For interested AIs, please contact Mr. John Chiu or Ms. Joyce Ip of Banking Supervision Department at GenAI_sandbox@hkma.gov.hk.