5attack classes tested per model release
31high-risk systems with a security file
100%of models covered by adversarial regression tests
Home/Case Studies/European insurer builds Article 15 security evidence for its underwriting models ahead of December 2027
European insurer builds Article 15 security evidence for its underwriting models ahead of December 2027
The insurer mapped the AI Act's named attack classes to tests and controls for each high-risk underwriting and claims model, producing a security file per system that its notified body could audit.
Contribute
Spot an error, or know a record we should add? Members can propose additions and submit corrections.