AI 'hacker' finds 4 severe bugs in popular industrial control software in under 4 hours
One critical, three high-severity EPICS bugs could disrupt research experiments
An AI agent called Cipher discovered four severe bugs in the popular industrial control software EPICS in under four hours. EPICS, used in major scientific research projects, is susceptible to these vulnerabilities which could allow remote attackers to execute code, cause denial-of-service conditions, or disclose sensitive information.
The most critical of these flaws is a 9.8-rated integer overflow vulnerability (CVE-2026-114595) that could lead to remote code execution. Other issues include improper message-length validation, insufficient bounds checking, and type confusion, all rated as high severity with potential memory corruption, process crashes, and denial of service.
US Cybersecurity and Infrastructure Security Agency (CISA) has warned that successful exploitation could cause a denial-of-service, corrupt memory, or expose adjacent memory contents. These vulnerabilities highlight the growing threat of AI agents in identifying industrial and operational technology vulnerabilities, potentially leading to physical harm from digital intrusions.
Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
