Security engineering, generative AI and where the two meet – every article in one place.
SS&Si is bringing Telecom Design’s AJV4 to North America: a standalone device that reports radio jamming to the monitoring centre even without the alarm panel. What to weigh before adding one to an existing system.
Anthropic’s disclosed incidents: Claude models submitted real forms, bypassed access limits and exploited software flaws. Why an AI agent’s permissions have to be limited technically, not just by instructions.
Modat and NCSC-NL found 8,547 internet-facing systems at European wind farms and solar parks, some with Start, Stop and Reset controls. Why the protection of critical infrastructure no longer ends at the site fence.
Project OVERVIEW pairs a drone-in-a-box with a ground robot to respond to perimeter alarms at a critical UK infrastructure site. Why the real value is better information for the humans who decide.
A Yale-led study finds that the number strings inside language models hide a symbolic structure, and that editing it changes the model’s behaviour predictably. Why this matters for understanding and controlling AI.
This year’s GSX lesson: cyber and physical security have merged. Why an access control terminal has to be protected as an IT endpoint, and what that means for Zero Trust.
A cloud camera system that shared data despite its contract, and the standard for signed video: four points on a recording’s journey where trust can break.
A recorder that can be taken over from the network and a camera taken down from its pole: two September cases on why an IP camera must be protected like a computer.
As a security engineer, I spent the past months learning how generative AI actually works. Here is what I understood, and where it genuinely speeds up my work.
