Breaking

IBM executive urges fighting AI attacks with AI as X-Force logs 44% rise

June 24, 2026 at 08:02 EDT

Security
AI Agents
Sales & Business

Gaurav Agarwal, Vice President, Technology at IBM India & South Asia, argued that as cybercriminals use AI to automate attacks and exploit vulnerabilities at greater speed, organisations must respond with AI-powered defences. Centered on a "fight AI with AI" approach, he called for identity-centric defence and stronger governance of AI agents.

IBM · Autonomous Security

Fighting AI With AI: Enterprise Defense Enters the "AI vs AI" Era

IBM unveils a vendor-agnostic, multi-agent system that maps exploit paths, detects anomalies, and contains threats at machine speed — as attackers increasingly automate intrusions with AI.

$5B

Investment to provide clients with "clean software images" that strip out vulnerabilities

Coordinating AI agents at the core: ARGO, ADA & ATOM

Domains spanned: applications, identity, data, network & cloud

The Multi-Agent Defense Loop

Three interoperable agents coordinate to defend at machine speed

ARGO

Autonomous Risk Governance Orchestrator

Orchestrates risk governance across the system

→

ATOM

Autonomous Threat Operations Machine

Automates threat operations & maps exploit paths

→

ADA

Autonomous Defense Agents

Autonomous defense, response & threat containment

Insights feed back into governance & risk frameworks ↻

Why the model is changing

Perimeter security is fading — attackers increasingly walk in "through the front door" using valid credentials.

Cloud migration & remote work blur the network boundary

Identity becomes a new attack vector

Least privilege + continuous monitoring for AI agents

The promise

Systemic, machine-speed defense replaces fragmented tools — welcomed as a way to match the pace of AI-driven attacks. Hybrid-by-design and data sovereignty via Red Hat integration.

The caution

AI agents carry new risks — prompt injection, data exposure, bias. Human-in-the-loop oversight remains for critical decisions, and large-scale operational evidence is yet to accumulate.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.