PhishLumos System Detects Phishing Campaigns Before Expert Verification

Researchers at Tokyo Metropolitan University developed PhishLumos, a system that identifies phishing campaigns by mapping domains, IPs, and certificates. In tests, it detected malicious activity an average of 8 days prior to expert verification across 103 phishing campaigns. The system revealed 192,407 additional URLs from 600 starting links, with 92% later flagged as malicious by VirusTotal. Phishing remains a significant threat, with over 880,000 complaints reported to the FBI in 2023, resulting in losses exceeding $12.5 billion. The system addresses the limitations of existing defenses, which often react to individual links rather than the broader campaign. PhishLumos adapts to content-inaccessible cases, where phishing sites may appear benign or unresponsive to automated scanners. This innovation aims to enhance detection and prevention of phishing attacks, particularly affecting vulnerable populations like older adults.

Key Points: • PhishLumos detects phishing campaigns 8 days before expert verification. • The system uncovered 192,407 additional URLs from 600 initial links, with 92% flagged as malicious. • Phishing complaints exceeded 880,000 in 2023, with losses over $12.5 billion.