Agentic AI platforms are reshaping security operations in corporate IT environments where speed is the primary design goal. Critical infrastructure operates under different constraints. NERC CIP-007, IEC 62443-2-3, and NIST SP 800-82 all require documented human authorization before security changes reach production. In a new article, OPSWAT explains why AI belongs at the analysis layer in OT environments where downtime carries safety or financial consequences.
Why Do Autonomous AI Decisions Fail in OT?
A protection relay in a power grid cannot be restarted mid-cycle. A configuration change to a PLC on a manufacturing line cannot be rolled back in seconds. An automated remediation step that fires during a production run affects physical processes, not just servers. Corporate IT security platforms prioritize response speed. OT security teams face a different requirement: accountability when the system is wrong.
The failure pattern repeats across incidents. A patch deploys outside a maintenance window at a substation. A firewall rule changes during a production cycle. A configuration change propagates across a distributed OT network before any operator reviews it. In each case, an automated system acted without human authorization, and by the time someone noticed, the change had already reached production.
What Do NERC CIP and IEC 62443 Require?
NERC CIP-007, IEC 62443-2-3, and NIST SP 800-82 all mandate human accountability for security changes in critical infrastructure. Autonomous workflows can support these requirements when governance is built into the process. They become a liability when “the system decided” replaces accountable human authorization.
| Framework | Scope | Key Requirement |
| NERC CIP-007 | Bulk electric system | Documented assessment, testing evidence, deployment timelines |
| IEC 62443-2-3 | Industrial automation and control systems | Authorization responsibilities for patches and configuration changes |
| NIST SP 800-82 | ICS security | Risk assessment, validation testing, and stakeholder coordination before deployment |
When no accountable human authorization event exists in the audit log, auditors cannot identify who approved the action.
How Does a Compromised Autonomous Platform Scale an Attack?
Security management platforms hold privileged access to push configuration changes, deploy patches, and manage policies across hundreds of deployments. When that platform operates autonomously, the attack surface expands: a compromised system can execute actions across every connected deployment before a human operator recognizes the breach. The attacker inherits the platform’s execution authority.
In early 2026, three critical zero-day vulnerabilities in a widely deployed patch management platform allowed unauthenticated remote code execution across enterprise environments. CISA flagged these as Known Exploited Vulnerabilities requiring immediate remediation. A platform that requires human approval before executing changes limits this blast radius. Stolen credentials alone are unlikely to trigger automated changes at scale.
Where Should AI Sit in the OT Security Stack?
AI delivers the most value at the analysis layer: risk-ranked vulnerability triage, configuration drift detection, anomaly correlation, and prioritized findings for human review. CISA guidance highlights visibility gaps as a core challenge for critical infrastructure organizations managing distributed assets. AI addresses this by aggregating event data, correlating signals, and surfacing prioritized findings. The analyst decides, and AI helps the analyst decide faster with better information.
EPAM Systems, a global digital platform engineering firm with approximately 40,000 employees across 30 countries, secured more than 70,000 devices using OPSWAT’s My OPSWAT Central Management and MetaDefender Endpoint platforms. The security team validated device compliance, detected unwanted applications, and identified unpatched vulnerabilities. EPAM also processed more than 50 million files per day at peak volume through MetaDefender Cloud. Every decision stayed with the security team.
Read the full article on OPSWAT.com.
Related from IIoT World:
