Cybercriminals Shift to AI-Powered Precision Attacks as Phishing Volume Drops 20% — Zscaler, Inc.

Event summary

Phishing volume fell 20% for the second consecutive year as attackers pivot to high-fidelity, AI-generated lures.
Services sector saw a 65.5% surge in targeted attacks, exploiting trust-based workflows like billing and renewals.
95.2% of phishing attempts now hide in encrypted traffic, bypassing legacy security stacks lacking deep TLS inspection.
ThreatLabz identified 413,524 AI-generated phishing sites, with 10% explicitly malicious.
Deception telemetry recorded 89.9 million hostile interactions from 1.37 million unique attacker IPs in six months.

The big picture

The decline in phishing volume signals a strategic shift in cybercrime economics, with attackers favoring quality over quantity. AI-powered tools are enabling faster, more precise attacks that evade traditional security measures. The surge in encrypted phishing attempts underscores the need for deep TLS inspection and Zero Trust frameworks to mitigate risks. This evolution in attack methods demands a reevaluation of cybersecurity strategies across industries, particularly in high-value sectors like government and manufacturing.

What we're watching

AI Weaponization: How the pace of AI adoption by cybercriminals will outstrip defensive capabilities.
Encryption Blind Spots: Whether organizations can adapt quickly enough to inspect 95% of encrypted traffic.
Zero Trust Adoption: The extent to which enterprises will prioritize Zero Trust architectures to counter evolving threats.