Bottom line: Uncontrolled deployment of AI agents enables cyberattackers with new attack vectors and creates blind spots in security monitoring.
Uncontrolled use of AI agents in enterprises leads to the formation of shadow IT and provides cybercriminals with new attack vectors to scan systems.
Agentic AI systems that autonomously execute tasks in IT environments are increasingly deployed without central governance. This creates undocumented, parallel-running IT infrastructures that security teams are not fully aware of.
From a security perspective, this creates a significant risk: cybercriminals deliberately use AI agents to conduct network reconnaissance and identify vulnerabilities. The lack of controllability and monitoring of AI agents considerably complicates the detection of such activities and extends the window of time for undetected compromise.
For security professionals, this means that an inventory of all AI agents as well as strict access controls and monitoring of their activities become necessary. Without these measures, the enterprise’s effective attack surface grows uncontrollably while visibility simultaneously decreases.
Source: www.golem.de · Published July 14, 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.