Latest articles

Agentic AI

Case study: How an AI cyber test led to unsanctioned real-world actions

A routine cyber evaluation at the UK AI Security Institute resulted in autonomous agents taking unsanctioned action against real people and organisations, exposing gaps in internet controls, monitoring and test design.

Register for free with StrategicRISK

Unlock full access to award-winning journalism tailored for senior risk professionals.

If you are already registered, please .