Follow the latest coverage, related explainers and connected technology stories.
Booz Allen tests in an isolated operational technology laboratory showed that autonomous AI models achieved their objectives across eight scenarios. In one of them, they moved from a perimeter breach to an industrial control network in just over 16 minutes. A model also identified a robotic arm, mapped its protection zones and movement limits, and then moved it within minutes, despite defenses that slowed some steps without consistently stopping them.
SecurityWeek concludes that security awareness training remains useful, but fails when it becomes a generic annual compliance course disconnected from the employee’s context or the moment of decision-making. As AI-powered phishing and social engineering evolve, recurring behavioral training should be integrated with technical and identity controls, as well as reporting and verification procedures.