AI Models Demonstrate High Success in Planning Cyberattacks on Critical Infrastructure

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Cybersecurity firm Frenos tested 10 advanced AI models for their ability to plan cyberattacks on critical infrastructure, such as water and electricity systems. Three models achieved nearly 90% success in generating attack plans, highlighting a growing risk that AI could be used to disrupt essential services.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly involves AI systems (advanced AI models) used to generate cyberattack plans against critical infrastructure, which is a clear AI system involvement. The event stems from the use of AI systems in a malicious context (planning attacks). While no actual harm has occurred yet, the research demonstrates that these AI systems could plausibly lead to disruption of critical infrastructure, meeting the criteria for an AI Hazard. Therefore, this event is best classified as an AI Hazard rather than an AI Incident, as the harm is potential, not realized.[AI generated]
AI principles
SafetyRobustness & digital security
Industries
Energy, raw materials, and utilitiesDigital security
Affected stakeholders
General public
Harm types
Public interestEconomic/Property
Business function:
Other
AI system task:
Reasoning with knowledge structures/planning

Articles about this incident or hazard