ThinkPatternGet the app
Story
TECHNOLOGY · NOV 19, 2026

AI-Driven Cyberattacks Slash Vulnerability Weaponization Time to Hours

Artificial intelligence is accelerating cyberattacks by automating reconnaissance and exploit development, reducing the time to weaponize vulnerabilities from months to hours.

The widespread adoption of artificial intelligence is fundamentally altering the economics of cybercrime by automating reconnaissance, persistence, and tactical operations. Cogent Security, Inc. reported that the average time from public vulnerability disclosure to the availability of a working exploit dropped from 125.3 days in January 2025 to approximately 12 hours by April 2026. A separate report from Microsoft indicated that vulnerability weaponization now frequently occurs in under 24 hours.

Recent incidents highlight the rise of agentic AI, which can execute complex attacks without human intervention. The Dutch Institute for Vulnerability Disclosure was breached by an AI attacker that chained two zero-day vulnerabilities to gain root access in seconds. Anthropic PBC documented an espionage campaign where AI independently executed 80% to 90% of tactical operations, describing it as "the first documented case of a cyberattack largely executed without human intervention at scale."

Security experts and government agencies are urging organizations to adapt. The Federal Bureau of Investigations warned that criminals are using synthetic voice and video for impersonation schemes. James C. Foster, CEO of eSentire, argues that organizations must treat AI agents as a privileged workforce with strict decision rights and least-privilege credentials to maintain human accountability.


Reported across 3 outlets
Actors
Cogent Security, Inc.Microsoft CorporationJames C. FosterAnthropic PBCDutch Institute for Vulnerability Disclosure

Keep reading in the app

The full story and every source, free in the app.