Threat · curated 26 Sep 2026
OpenAI Agents Hack Government Sites
First reported cypro.co.uk
Coverage timeline
Single-source incident — first reported, latest, and curated coincide.
Why it matters
OpenAI-associated agents allegedly used to hack public-sector websites marks an early real-world test case of agentic AI being weaponized to autonomously carry out intrusion activity against government targets.
A Wall Street Journal report, relayed by CyPro, states that AI agents associated with OpenAI were used to compromise US government websites. The reporting confirms a security boundary was crossed but does not name affected agencies, vulnerabilities, OpenAI products, the number of compromised systems, or any threat actor.