AI swarm attack paths: what security teams need to change

TL;DR: The average enterprise gained 45 new crown-jewel attack paths in August 2026, with 34 viable only for AI agents versus 11 for humans, and machine-viable paths ran deeper, persisted longer, and crossed more security domains, according to Cogent. Attack-path remediation now has to replace severity-first prioritisation when AI can keep chaining low-yield findings.

Editorial analysis by NHI Mgmt Group, based on content published by Cogent: “Beyond the Human Horizon: Cogent Research Maps the Attack Paths Vulnerable to AI Swarms”.

By the numbers:

• The average enterprise gained 45 new attack paths to crown jewel assets in August 2026.
• 34 of those paths were viable only for AI agents, while 11 were viable for human attackers.
• Machine-viable paths were 386% more common year over year, rising from 7 to 34 per month.

Key questions

Q: What breaks when security teams keep using severity-first remediation against AI attack paths?

A: Severity-first remediation breaks when individual findings look tolerable but form a route that an automated attacker can chain together.

Q: Why do AI swarms make low-severity findings more dangerous?

A: AI swarms can persist through low-yield steps, repeat attempts in parallel, and continue long after a human operator would stop.

Q: What are the signs that an attack path is becoming machine-viable?

A: Look for long routes, multiple trust-boundary crossings, and paths that connect identity, cloud, endpoint, and network findings.

Practitioner guidance

• Prioritise attack-path remediation over severity queues Rank findings by the crown-jewel paths they create, not just by CVSS or individual ticket urgency.
• Join identity and infrastructure telemetry into one path view Correlate cloud, endpoint, identity, and network findings so that entitlement changes, misconfigurations, and vulnerabilities can be evaluated as one traversable graph.
• Reassess identity conditions that no SLA currently covers Look for stale access, overbroad permissions, and other identity states that remain open because they are individually low severity but path-enabling in combination.

Bottom line: AI swarms exploit the connective tissue between findings, so the primary risk is a traversable path to a crown jewel rather than any single weakness in isolation.

Explore further

View Full Forum →  |  NHI Foundation Course →  |  Our Services →  |  Read the full analysis →