IBM Launches Autonomous Security to Combat Agentic AI Attacks

Starting April 15, 2026, IBM launched a suite of cybersecurity measures designed to protect enterprises from agentic attacks fueled by frontier AI models. These threats can autonomously discover vulnerabilities and accelerate the attack lifecycle at machine speed, which the company says renders manual security processes insufficient.

The strategy consists of two primary components. IBM Consulting is offering a new cybersecurity assessment to help organizations identify security gaps and AI-specific exposures. Simultaneously, the company introduced IBM Autonomous Security, a multi-agent-powered service that employs vendor-agnostic digital workers to coordinate decision-making and response across a client's existing security stack.

Mark Hughes, Global Managing Partner of Cybersecurity Services at IBM Consulting, characterized these frontier models as a new category of enterprise threat that is systemic and increasingly autonomous. He stated that AI-powered offense requires an AI-powered defense to ensure organizational resilience matches the speed of these autonomous threats.