The Hong Kong Insurance Authority issues a white paper: Alliance-style learning to enhance the operational efficiency and data privacy of the insurance industry.

On November 3rd, the Insurance Authority of Hong Kong (IA) announced the release of the white paper "Federated Learning: Opening the Door to Innovation in the Insurance Industry" during the Hong Kong Fintech Week. Mr. Clement Cheung, the Chief Executive of IA, stated, "In the face of rapidly changing market conditions, financial regulators must maintain a balanced and open-minded attitude to promote market-compatible and responsible innovation. Given this background, federated learning plays an important role in facilitating cross-boundary data collaboration while safeguarding personal privacy." The white paper outlines that the federated learning research program will commence in March 2023. Federated learning is an advanced machine learning technology that allows models to be trained on decentralized data sets. Unlike traditional machine learning that requires data sets to be centralized on a single server, federated learning enables models to be trained directly on the devices or servers where the data is located, ensuring data privacy protection and compliance with regulations, and promoting cross-institutional collaboration. IA pointed out that for the insurance industry dealing with large amounts of sensitive customer data, federated learning can improve operational efficiency and data privacy. IA stated that the federated learning research program does not utilize an open-source framework but has developed a platform specifically tailored to the needs of the insurance industry, incorporating advanced privacy-enhancing technologies, optimization algorithms, and a robust modular architecture to enhance security, efficiency, and flexibility. Evaluation results indicate that the platform can effectively enhance data protection levels and model performance. IA emphasized that the results highlight the multiple advantages of federated learning, providing innovative strategic opportunities for insurance companies and opening up new business opportunities. These advantages include the ability to integrate diverse data sources to develop smarter predictive models without compromising data privacy, enhance the accuracy of claim and customer behavior predictions, support optimized pricing, resource allocation, and marketing strategies; enable institutions to jointly train models without sharing sensitive data, promote secure cross-industry collaboration, help break down data silos and regulatory barriers, generate richer insights, and more robust models; and promote responsible AI practices in line with decentralized data and evolving regulatory standards, enhancing data privacy and strengthening customer trust in data usage. In his keynote speech, Clement Cheung also highlighted a series of initiatives led by IA to promote advanced technology applications and strengthen the operational resilience of the insurance industry, including the Open Application Programming Interface Framework, the Cybersecurity Assessment Framework, the Artificial Intelligence Acceleration Scheme, and the federated learning white paper. Additionally, the Insurance Authority of Hong Kong and Invest Hong Kong co-hosted the "InsurTech Forum," featuring interactive discussions on how technological innovations reshape the insurance industry's market promotion strategies, corporate culture, and customer experience.