Data plays a crucial role in the insurance sector and in digital transformation. However, data privacy concerns and strict data handling regulations are hindering the development of data driven solutions and AI innovations in the industry. Federated learning (FL) offers a promising solution to this problem, as it allows insurers to leverage machine learning (ML) while safeguarding individual data privacy. Against this background, in March 2023 the Insurance Authority (IA) and the Hong Kong Applied Science and Technology Research Institute (ASTRI) undertook this research project with the aim of exploring some potential FL applications for the insurance industry.
Privacy-Preserving AI Collaboration: Enables insurers to train advanced models on decentralised data without sharing sensitive information, overcoming privacy and regulatory challenges.
ASTRI’s Custom Platform: A tailored, ready-to-market Federated Learning solution with robust privacy features, optimised for insurance use cases.
Proven Results: Three PoCs demonstrated improved customer propensity prediction, claims forecasting, and renewal modelling through secure cross-sector data collaboration.
Actionable Insights: Practical guidance on risk management, compliance, and adoption strategies to drive responsible innovation in Hong Kong’s insurance sector.
