AI Models Used in Cyber-Attack on Water and Drainage Facility

Researchers at Dragos reported that commercial AI models from OpenAI and Anthropic were used to assist in planning and executing a cyber-attack targeting the operational technology (OT) of a water and drainage facility. The attack involved leveraging large language models (LLMs) to support malicious activities against critical infrastructure. No specific dates, technical indicators, or CVE IDs were disclosed in the findings. The incident highlights the use of AI tools in cyber-attacks on essential services, though further details about the facility’s location or the extent of the impact remain undisclosed.