Evidence · DiggingBeagle record

Phishing in 2026: AI and Deepfakes Changed the Attacks, but Not the Defensive Basics

This Habr interview with BI.ZONE AntiFraud and Digital Risk Protection staff is useful for the defensive problem that remains after generative AI improves impersonation. The interview describes more personalized phishing, synthetic voice and video, multi-stage social engineering, TOAD and phishing-as-a-service, then recommends verification through independent communication channels, two-factor authentication using authenticator applications, protected communications, deepfake-detection tooling, employee training and second-person approval for critical operations. Its quantitative statements need separate denominators: it attributes a 64% share of targeted attacks beginning with phishing email to Threat Zone 2026, reports 12,500 fraudulent sites detected in Q1 2026 versus 17,500 a year earlier, and gives a BI.ZONE estimate of more than RUB 18 billion stolen through phishing schemes in 2025. Those figures describe different populations and should not be combined into one trend or victim count. The controls also block different prerequisites: independent verification and second-person approval can interrupt a single-channel impersonation path, while two-factor authentication mainly hardens account access and does not by itself stop a victim from approving a fraudulent transfer. The interview provides no controlled effectiveness or false-positive data for deepfake detectors.

Published

Apr 30, 2026

Source role

secondary reporting