ASTRI and IA Unveil Hong Kong’s First White Paper on Federated Learning in the Insurance Industry

3 November 2025 – The Hong Kong Applied Science and Technology Research Institute (ASTRI) and the Insurance Authority (IA) have jointly released a White Paper “Federated Learning: Unlocking Innovation in the Insurance Sector” today, highlighting the transformative potential of Federated Learning (FL) for privacy-preserving data collaboration in the insurance industry.

The While Paper presents the findings of an Innovation and Technology Fund-supported research project that involved the development of a ready-to-market platform for testing FL applications. Through collaboration with insurers and cross-sector partners, the platform showcases how FL can revolutionise claims management, renewal predictability and customer insight analytics.

Driving Innovation Through Collaboration
Ir Dr Ted Suen, Chief Executive Officer of ASTRI, emphasised the role of FL in advancing data-driven decision-making: “ASTRI is honoured to collaborate with the Insurance Authority in advancing innovation within Hong Kong’s insurance ecosystem. FL is a game-changer in how organisations can collaborate securely while respecting data privacy. The White Paper and the FL platform developed by ASTRI chart a new course for Hong Kong’s insurers to lead the region in harnessing advanced technologies that deliver tangible value whilst building enduring trust with customers and stakeholders.”

Mr Clement Cheung, Chief Executive Officer of the Insurance Authority (IA), pointed out that “Given the rapidly evolving market landscape, financial regulators must navigate in a balanced and enlightened manner to promote inclusive and responsible innovation. Seen in this context, Federated Learning plays a useful role in enabling cross-sector data collaboration without compromising personal privacy.”

Three practical use cases were tested and have demonstrated the wide-ranging benefits of FL for insurers:

1. Customer Identification: Leveraging engagement insights to enhance AI models for identifying potential customers.
2. Claims Forecasting: Using clinical data to predict insurance claims probabilities.
3. Renewal Predictions: Analysing credit data to forecast customer renewal likelihoods.

To ensure responsible implementation, the White Paper recommends insurers adopt a comprehensive assessment framework addressing data protection, regulatory compliance, and ethical considerations. It also outlines a strategic roadmap to accelerate adoption, focusing on technical optimisation, organisational readiness, and ecosystem collaboration.

The White Paper is now available on the ASTRI’s corporate website, please visit: https://bit.ly/4hDcgDb