AI on Both Sides of the Wire: Building Cryptographic Guardrails for Resilient Network Operations

Artificial intelligence is rapidly becoming a core operational tool for network operators - powering fraud detection, security analytics, automation, and optimization at scale. At the same time, attackers are leveraging AI to automate reconnaissance, impersonation, malware generation, and supply chain attacks at machine speed. This creates a fundamental asymmetry: defenders must be correct every time, while attackers need only succeed once.

This session examines how critical infrastructure can remain resilient as AI accelerates both offense and defense. While AI driven analytics significantly improve detection, detection without enforcement increases alert volume and does not stop fast moving, adaptive attacks.

The session explains why cryptographic guardrails - strong device and workload identity, software and model provenance and integrity, secure boot and signing pipelines, attestation, and revocation - are essential to operationalizing AI insights. These trust primitives enable AI driven detections to be translated into enforceable, automated controls, allowing operators to isolate compromised devices, restrict untrusted workloads, and contain attacks at machine speed.