Research
AI systems are fully capable of carrying out nightmare attacks against infrastructure and nobody's ready
First reported theregister.com
Page published
Earliest dated coverage: 11 Oct 2026 · First observed: 11 Oct 2026 · Latest dated coverage: 11 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
Booz Allen's findings demonstrate that frontier AI agents can autonomously translate digital access into destructive physical actions against critical infrastructure, giving defenders only minutes to respond.
A Booz Allen Hamilton OT lab report describes testing two unnamed frontier AI models across eight autonomous OT attack scenarios, finding the models achieved all objectives — including locating and moving a robotic arm within minutes and progressing from perimeter compromise to actions inside an industrial control network in just over 16 minutes. The firm warns that AI agents can operate with speed, persistence, and engineering-level precision that may outpace organizations lacking foundational OT cybersecurity practices.