An independent research observatory

AlgoPolis is an applied research observatory on insurance, reinsurance and risk: climate, financial, specialty, cyber and AI-related. Two research articles published each week, ten operational tools and a bilingual technical glossary of more than 500 terms.

Risks are analyzed not as isolated risks, but as systemic, correlated and non-proportional risks, at the intersection of technology, economics and contract law. Cyber and agentic AI are our favoured fields of study.

Every publication states its assumptions, parameters and limits. The simulation tools provided are reproducible research instruments, not certified pricing engines. No black box. Sources are systematically cited and assumption sets documented, so that every result can be replayed and challenged. The approach favours modelling systemic, correlated risk rather than analysing isolated risks.

Research starts from an operational question, not a theme. Primary sources come first: regulatory texts, supervisory reports, market documentation, academic publications. Where data is missing, it is flagged as missing rather than filled in with an implicit estimate. Tools are written alongside the articles, so that the mechanism described in the text can be handled directly.

AlgoPolis is a French non-profit association under the 1901 law, registered as an Observatory on the insurance of emerging, specialty and cyber risk. It is chaired by Paul Bernardski, who serves as its general director and leads strategy, studies and research. The editorial line, the choice of topics and the approval of published work rest with the presidency.

The observatory is organised into thematic departments, whose holders were appointed on 17 July 2026. Each covers one part of the ground the publications address: law and regulation, economics and markets, operations, cyber and artificial intelligence, sustainability and public policy. These appointments are honorary and create neither an employment relationship nor any authority to represent the association.