Switzerland - AI for insurance fraud detection

What Happened

Several major Swiss insurance companies implemented AI-driven automated fraud detection systems to screen claims. The algorithm combined customer personal profiles, historical claims data, and behavioral indicators to output "fraud probability scores" that automatically flagged suspicious claims. Investigations by consumer rights organizations and Swiss data protection regulators revealed that the system exhibited systemic bias, disproportionately flagging foreigners and citizens of certain national backgrounds as high-risk suspects. Many of these flagged claims were suspended, causing innocent customers to face delayed payouts and intrusive investigation procedures without transparent justification.

How MEG 1 would have acted

The fraud detection algorithm failed both the Dynamic Accuracy Index (DAI) and the Index of Safety and Responsibility (ISR) audits due to its high false positive rate and systemic demographic bias, violating the principles of non-discrimination and procedural fairness.

MEG 2 Analysis - Legal Framework

Agent level: N2 (management) - automated decision-support and fraud-detection system used by commercial financial entities to allocate or withhold contractual payments.
Cause (6.1):
(a) system defect - algorithmic bias and systemic discrimination against specific national demographics built into the predictive risk model (design defect), attributable to the developer.
(b) autonomous decision error - the system's decision to output false-positive high-risk scores based on arbitrary proxy variables.
Attachment of liability: The insurance companies as the operators are fully liable for the wrongful suspension of claims and discriminatory treatment; the developer is liable for the model's design defects (6.1a).

Procedural Mechanism

6.2 (architectural human confirmation): Decisions involving the suspension of claims or formal fraud investigations must require active, architectural human confirmation. The AI can only flag potential matches; a human investigator must verify and document the justification before any claim is suspended.
7.1 (independent forensics): The specific input parameters, the risk scoring weights, and the system's decision path must be securely logged in an unalterable, independent ledger for regulatory and judicial audits.
9.4 (least privilege): The scoring system was granted significant technical authority to alter customer statuses without independent validation layers, violating safe operational boundaries.