An Empirical Study of Modern Anti-Phishing Defenses

Phishing remains one of the most effective initial access vectors despite continuous advancements in platform-level security. This paper presents an empirical evaluation of modern anti-phishing defenses by comparing adversary-in-the-middle (AITM) phishing frameworks, automated one-time password (OTP) relay techniques, and legacy credential-harvesting phishing methods. Experiments were conducted in a controlled and authorized laboratory environment using test accounts and self-owned domains. Results show that modern platforms effectively neutralize MITM-based attacks through layered defenses including TLS fingerprinting, IP reputation analysis, behavioral anomaly detection, and contextual multi-factor authentication. In contrast, legacy phishing techniques continue to succeed primarily due to human susceptibility. The findings highlight a shift in phishing viability and emphasize the importance of user-centric defenses alongside technical controls.