Recently, the application of Artificial Intelligence (AI) in the realm of financial services has been steadily increasing, particularly with the rapid development of generative AI, influencing a wide range of areas. In order to assist financial institutions in leveraging the benefits of AI technology, and to effectively manage risks, ensure fairness, protect consumer rights, maintain system security, and achieve sustainable development, the Financial Supervisory Commission (FSC) has formulated a draft document based on the Executive Yuan's "Taiwan AI Action Plan 2.0" and the promotional strategies discussed at the Executive Yuan's "Special Meeting for Coordination of Digital Policies and Legislation," and by referencing AI guidelines issued by financial regulators of major countries and international organizations. This draft, also combined with the current state and supervisory policy direction of Taiwan's financial market, is titled "Proposals for Core Principles and Associated Policy for AI Application in the Financial Industry" (hereinafter referred to as the draft). Today, the FSC has published this draft on the public policy participation platform "JOIN" to solicit external feedback. All stakeholders are welcome to provide their suggestions on the platform within the next 10 days.
The draft sets forth six core principles for AI application in financial sector, as follows:
1.Establishing governance and accountability mechanisms: This principle emphasizes that financial institutions should be responsible for internal governance and protection of consumer rights when using AI systems, and exercise appropriate supervision over AI system risk management and usage.
2.Emphasizing fairness and human-centric values: This principle highlights that financial institutions should focus on offering fair and inclusive financial services when utilizing AI systems, in line with human-centric and controllable principles.
3.Safeguarding privacy and customer rights: This principle stresses that financial institutions should fully respect and protect customer privacy when using customer data, and effectively manage and leverage the relevant information to enhance consumer confidence and satisfaction.
4.Ensuring system robustness and security: This principle emphasizes that financial institutions should commit to maintain the robustness and security of AI systems, and exercise proper risk management and oversight over third-party providers to offer better financial services to consumers.
5.Emphasizing transparency and explainability: This principle underscores that financial institutions should ensure transparency and explainability in AI system operations, and should appropriately disclose relevant information when using generative AI for business operations or as auxiliary tools for financial services.
6.Promoting sustainable development: This principle highlights that financial institutions should ensure their development strategies and implementations align with sustainable development principles when using AI systems, and make efforts to protect the right to work for employees.
In recent years, the rapid development of generative AI has not only helped improve production efficiency and provide diverse services, but it has also raised issues related to personal data protection and information security. The FSC has been closely monitoring the progress of generative AI and, referring to the draft guidelines of the National Science Council on "Guidance for the Executive Yuan and its Affiliated Agencies on the Use of Generative AI," has listed considerations in the draft. These include recommendations that financial institutions should not solely rely on information produced by generative AI for decision-making or judgment, and financial institution personnel should conduct objective and professional final assessments of outputs generated by generative AI regarding their associated risks.
Furthermore, based on the announced draft, the FSC plans to launch a comprehensive set of AI-related policies in response to AI development. In addition to providing guidelines for the financial industry's use of AI based on the six core principles, the FSC also intends to review and adjust relevant regulations as needed, employ AI technologies in suptech, engage in exchanges and collaborations with international organizations and foreign financial regulatory agencies, encourage the financial sector to actively participate in AI research, development, and application, and to introduce best practices. The FSC will also assess the actual state of AI application in the financial industry, mandate financial industry associations to establish self-regulatory standards and best practices for using AI systems, oversee financial institutions in ensuring fair treatment of customers and adherence to friendly financial guidelines, and reduce the digital divide through financial knowledge promotion activities.
The FSC stated that in formulating this draft, numerous internal and external meetings have been convened, taking into account the opinions of relevant departments, experts, scholars, and various financial industry associations. To further refine and align the principles and policies with practical operations, the draft has been released on the "JOIN" platform to extensively gather external feedback. After considering this feedback, the finalized version will be officially released for financial institutions to follow.
Moving forward, the FSC will continue to monitor the challenges and opportunities faced by the financial industry in integrating AI technology, as well as urge financial institutions to strengthen risk management, consumer protection, information security, and digital rights for vulnerable groups to develop trustworthy AI. Based on the core philosophy of responsible innovation, the FSC will guide the financial industry to adopt technological innovations, enhance service efficiency and quality, achieve the objective of financial inclusion and at the same time balance consumer rights, ethics, and financial market order.