AI Accelerates Cyberattacks, Cutting Breach Times to Hours

Artificial intelligence is reshaping the cyber‑threat landscape. Large‑language models enable highly targeted, realistic spear‑phishing messages and deep‑fake audio or video that fool employees and bypass traditional authentication. Open, inexpensive models such as GLM‑5.2 lower the barrier for attackers, allowing automated reconnaissance and multi‑step attacks at scale. CDNetworks reports that AI‑driven bots generated 64 million malicious requests per day in 2025, turning high‑volume API traffic into a primary attack vector and causing billions of blocked API requests each month. Palo Alto Networks warned that fully automated breach cycles can now be completed in roughly 72 minutes, a dramatic contraction from days‑long operations in the past. Security Operations Centers are adapting by embedding AI throughout investigation workflows, letting analysts correlate alerts and telemetry at machine speed while focusing on high‑value decisions. The combined effect is a rapid compression of the attack timeline and a growing need for organizations to assume breach is possible and build resilient, AI‑aware defenses.