A false positive occurs when an alert or classification identifies a condition that is not actually present. In payment screening, it means rejecting a genuine customer transaction as fraudulent. Reliable systems must detect real fraud while minimizing these mistaken rejections. section 11.11
A legitimate payment rejected
CMSPI’s analysis found that one in five rejected online transactions was a false positive. More than half of the affected customers then took their business elsewhere, turning an anti-fraud safeguard into lost customers and revenue. section 11.11
False positives also burden investigations. Anti-money-laundering systems built around predefined rules can generate large numbers of false alarms, reportedly consuming as much as 90% of compliance analysts’ time. Unsupervised learning and clustering can help expose unusual transaction patterns, although evaluating such models remains difficult without labeled outputs. section 3.6
