Security operations across Asia are moving from monitoring towards autonomy. AI now identifies threats, triages alarms and increasingly influences operational responses, from access control decisions to incident escalation and responder dispatch. Meanwhile, the security infrastructure it operates within—including access control, video analytics and building management systems—has itself become an expanding attack surface.

The critical question is no longer simply whether to adopt AI, but how to govern systems capable of making decisions with real-world physical consequences—and whether organisations can reliably intervene when those systems behave unexpectedly.

Drawing on 26 years of experience securing critical infrastructure, operational technology and regulated environments, Leonard examines the emerging challenges of AI-enabled physical security and the governance principles needed to maintain human accountability and operational control.

He presents five practical controls: defining AI decision boundaries, identifying untrusted inputs, managing AI supplier dependencies, establishing human accountability for consequential physical actions, and validating mechanisms to suspend or override autonomous operations.

Attendees will leave with five questions every organisation should ask before trusting AI-enabled security systems.