保险业人工智能监管：平衡消费者保护与创新

REGULATION OF ARTIFICIALINTELLIGENCE IN INSURANCE:Balancing consumer protectionand innovationDennis NoordhoekDirector Public Policy & Regulation, The Geneva Association The Geneva AssociationThe Geneva Association was created in 1973 and is the only global association of insurancecompanies; our members are insurance and reinsurance Chief Executive Officers (CEOs).Based on rigorous research conducted in collaboration with our members, academicinstitutions and multilateral organisations, our mission is to identify and investigate keytrends that are likely to shape or impact the insurance industry in the future, highlightingwhat is at stake for the industry; develop recommendations for the industry and forpolicymakers; provide a platform to our members and other stakeholders to discuss thesetrends and recommendations; and reach out to global opinion leaders and influentialorganisations to highlight the positive contributions of insurance to better understandingrisks and to building resilient and prosperous economies and societies, and thus a moresustainable world.Geneva Association publications:Pamela Corn, Director CommunicationsHannah Dean, Editor & Content ManagerKelly Gailey, Communications & Events ManagerSuggested citation: The Geneva Association. 2023.Regulation of Artificial Intelligence in Insurance: Balancing consumer protection and innovationAuthor: Dennis Noordhoek. September.© The Geneva Association, 2023 All rights reservedwww.genevaassociation.orgPhoto credits:Cover page – Mahdis Mousavi on unsplash 2 ForewordExecutive summary1.Introduction2.Defining AI2.1 Definitions2.2 Advanced analytics vs AI2.3 Machine learning2.4 Developing existing definitions3.Benefits of AI in insurance3.1 How AI is transforming insurance3.2 Benefits of AI4.Zoom in on AI and underwriting4.1 The underwriting process and AI: What is new?4.2 Changes to decision-making: AI vs humans5.Risks and concerns related to AI5.1 Transparency and explainability5.2 Bias, indirect discrimination and risks to fairness5.3 Data input5.4 Excessive price differentiation and exclusion5.5 How insurers are addressing and mitigating AI-related risks and concerns5.6 Are AI risks new to insurance?5.7 AI risks and the business model of insurance6.Regulation of AI6.1 Jurisdiction-specific developments6.2 How existing regulation addresses AI risks6.3 Balancing innovation with customer protection7.Conclusions and recommendationsReferencesContents 5679101010111213141617182021212223232626272831343537 ACKNOWLEDGEMENTSWe would like to thank the members of The Geneva Association (GA)’s Public Policy andRegulatory Working Group for reviewing the report and for the advice and feedback theyprovided.We also greatly benefited from interviews and discussions with the following experts:•Chubb•Julian Arevalo (EIOPA)•Lucy Ashton (ABI)•Birny Birnbaum (Center for Economic Justice)•Commissioner Kathleen A. Birrane (Maryland Insurance Administration)•Alessandro Bonaita (Generali)•Matt Brewis, Karen Croxson (U.K. FCA)•Chris Dolman (IAG)•Michael Drobac (AIG)•Michael Foehner, Tobias Wassmann (Swiss Re)•Oliver Goodenough (Vermont Law School)•Arthur Hilliard, Danilo Gattullo (Insurance Europe)•Joe Jones (International Association of Privacy Professionals)•Xuchun Li (Monetary Authority of Singapore)•Amedeo Santosuosso (Scuola Universitaria Superiore IUSS Pavia)•Dave Snyder (American Property Casualty Insurance Association)•Anika Stehr, Julia Perl (Hannover Re)•Lutz Wilhelmy (Actuarial Association of Europe)•Leigh Wolfrom (OECD)•Jing Xiao (Ping An) 4 Foreword The expansion of the use of artificial intelligence (AI) in insurance is emblematic of thebroader digital revolution sweeping across industries worldwide. AI has the potential toprofoundly transform the insurance industry, as AI algorithms factor into risk assessment,claims processing and customer interactions.There are unprecedented opportunities emerging, such as more affordable insurance prod-ucts and more satisfied insurance customers – and an overall enhanced value propositionfor the insurance industry.However, AI also poses challenges to upholding the core principles of fairness, transparencyand accountability.In navigating this territory, it is essential to approach the regulation of AI in insurance witha commitment to both enabling innovation and safeguarding the interests of all stake-holders. This report places insurers’ use of AI and its potential risks in the broader, dynamiclandscape of regulatory frameworks across different jurisdictions. It further assesses if andhow regulation addresses risks such as bias and discrimination.The regulation of AI in insurance is about more than creating rules and standards. It isabout fostering a culture of responsible innovation; it is about promoting inclusivity,empowering customers and preserving trust. We hope the guidance we offer to regulators,policymakers and insurers supports the collaboration and alignment required for AI to ben-efit all who rely on insuranc