基于风险监管的人工智能：又一个“锦上添花”的工具，还是颠覆者？

AI FOR RISK-BASEDSUPERVISION Another “Nice to Have” Tool or aGame-ChangerPublic Disclosure Authorized Matei Dohotaru, Marin Prisacaru, Ji Ho Shin and Yasemin Palta AI FOR RISK-BASEDSUPERVISION Another “Nice to Have” Tool or aGame-Changer January 2025 © 2025 International Bank for Reconstruction and Development / The World Bank1818 H Street NWWashington DC 20433Telephone: 202-473-1000 This work is a product of the staff of The World Bank with external contributions.The findings, interpretations, and conclusions expressed in this work do not necessarily The World Bank does not guarantee the accuracy, completeness, or currency of the dataincluded in this work and does not assume responsibility for any errors, omissions, ordiscrepancies in the information, or liability with respect to the use of or failure to usethe information, methods, processes, or conclusions set forth. The boundaries, colors, Nothing herein shall constitute or be construed or considered to be a limitation upon orwaiver of the privileges and immunities of The World Bank, all of which are specifically Rights and Permissions The material in this work is subject to copyright. Because The World Bank encouragesdissemination of its knowledge, this work may be reproduced, in whole or in part, for Any queries on rights and licenses, including subsidiary rights, should be addressed toWorld Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC Cover image: Veronica Gadea, GCS, World Bank Group.Accessibility: Will Kemp, GCS, World Bank Group. Contents AbbreviationsviAcknowledgementsviiExecutive summary1Main challenges faced by the financial sector supervisors3Empowering financial supervisors with AI capabilities14Use case of AI in supporting activities of the supervisory authorities22AI related risks and concerns28AI implementation challenges32AI adoption strategies39Looking into the future43Annex 1- ML & AI enterprise reference architecture50 Abbreviations AIArtificial IntelligenceASICAustralian Securities and Investments CommissionASXAustralian Securities ExchangeECBEuropean Central BankMAIThe Market Analysis and IntelligenceLLMsLarge Language ModelsAML/CFTAnti-Money Laundering and Combating the Financing of TerrorismFSAPFinancial Sector Assessment Program Acknowledgements This Paper was written by the Financial Stability and Integrity Unit (EFNFS), partof the World Bank’s Finance, Competitiveness, and Investment (FCI) GlobalPractice, and ITS (Information and Technology Solutions) Technology and Peer reviewers: 1) Harish Natarajan, Practice Manager, EFNFI; 2) Gunhild Berg,Lead Financial Sector Specialist, EECF2; 3) Olushola Ibironke Joanne Martins, ITOfficer, Business Management II, ITSIO; 4) Mert Ozdag, IT Officer, Engineering I,ITSIO; 5) Francesca Hopwood Road, Centre Head, BIS Innovation Hub – London; Please note that some sections of the paper have been drafted using World BankGroup approved generative AI technology. The Bank confirms that relevant sections Executive Summary Regardless of our individual perspectives on Artificial Intelligence (AI), it couldtransform our personal and professional lives at an unprecedented pace. It willalso impact one of the most regulated and supervised industries in the world–the financial sector. Risk-based supervision (RBS) has been the gold standardfor financial sector supervision over the past two decades, promising to aid The remarkable advancement of AI in recent years, both in terms ofperformance and accessibility, promises to revolutionize numerous industries,including the financial sector. Although attempts to use AI for supervisorypurposes began some years ago, they were mainly experimental and limitedto a few high-capacity supervisory authorities. However, recently, thanks to Initial attempts by global supervisors to utilize AI for various supervisorypurposes have yielded promising results. Areas that have traditionally sufferedfrom a lack of sufficient resources, such as Anti-Money Laundering andCombating the Financing of Terrorism (AML/CFT), consumer protection, credit The benefits of AI for financial sector supervision extend beyond the automationof some manual activities. In the near future, we are likely to see AI end-to-end automated supervisory processes, particularly those that do not require Of course, all these benefits do not come without risks and costs for supervisoryauthorities. The reengineering of supervisory processes based on AI capabilitieswill require effort but, more important, it will require the readiness to deal withnew risks generated by the use of AI functionalities. Preparing for the transition There is no expectation that AI will replace humans in the area of financial sectorregulation and supervision, even in the long term. Humans will continue to bedecision-makers on all critical aspects related to the supervision of the financialsector. However, we foresee an increasing symbiosis between humans and AI in Recognizing that t