Insurers, pension funds and the actuarial tradition
Read the PDF edition (2 pages)
Basel's insurance industry, which Part I described, has a long tradition of quantitative modelling. The actuarial profession was among the first to apply statistics to commercial decisions. The Swiss Solvency Test, for its part, requires insurers to model their risks in a way that anticipated many of the governance questions that artificial intelligence now raises.1 The merged group formed by Helvetia and Baloise, whose combination was completed in December 2025, and the other insurers present in Basel use machine learning in underwriting, pricing and claims handling.2
In underwriting, models assess risks from data that would once have required a site visit or a medical examination. Property insurers estimate the exposure of a building to flood or hail from satellite images and geographic data, while motor insurers price policies from vehicle and driver data. In life and health insurance, the use of such methods is more sensitive, because it touches on health data and on the principle of solidarity between policyholders. Claims handling is the second main field. Insurers use models to triage claims, to settle simple cases automatically and to identify suspicious claims for investigation, which reduces costs and shortens the time a policyholder waits for payment.3
Reinsurers and the captives that Part I, Chapter V described face a further question. Artificial intelligence creates new risks that must themselves be insured, such as the liability of a firm whose model causes damage, errors in automated advice and losses from AI-enabled cyber attacks. Specialist products for such risks began to appear in the London and Bermuda markets after 2023. Their development will depend on whether insurers can model losses for which there is little historical experience.4 Pension funds and investment foundations, described in Part I, Chapter IV, use artificial intelligence mainly through their asset managers and administrators, which makes the oversight of outsourced services a central governance task for their boards.
Notes
- On the Swiss Solvency Test and the internal models of insurers, see Part I, Chapter II; Arts. 9 and 46 of the Insurance Supervision Act (ISA, SR 961.01) and the Insurance Supervision Ordinance (ISO, SR 961.011). To be verified. ↩
- On the merger of Helvetia and Baloise, see Part I, Chapter I. On the use of AI by Swiss insurers, see the publications of the Swiss Insurance Association (SIA). Status to be verified. ↩
- On AI in underwriting and claims, see the reports of the European Insurance and Occupational Pensions Authority (EIOPA), in particular its Artificial Intelligence Governance Principles (2021) and its opinion on AI governance and risk management (2025), and the publications of the Swiss Insurance Association. Status to be verified. ↩
- On insurance products covering AI-related liability and performance risks, see market reports of the large reinsurers and the Lloyd's market since 2023. Status to be verified. ↩
For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of August 2025.