Governing the Invisible: Securing Machine Identities and Autonomous AI Agents in Modern Defence

A techUK roundtable on a problem defence has quietly acquired: non-human identities such as machine workloads, microservices and system tokens now outnumber human operators by more than 100 to 1, and they are the fastest-growing lateral attack surface in military systems. Automated decision-making is already embedded in logistics, threat analysis and signal processing. This is a working discussion rather than a panel. Registration and networking from 09:30, roundtable 10:30 to 12:00, networking to 12:30. Two questions anchor it: what safeguards and real-time kill-switches keep humans in control of autonomous decision-makers, and how hard-coded tokens get rotated across mission software pipelines and tactical cloud. Sponsored by Palo Alto Networks. Central London, places are limited, and the exact address is sent to approved attendees. Why go: a short, candid room where defence and industry argue about agent control before policy settles it. Best for security architects and anyone selling autonomy into defence.