Every day, we hear more instances of AI agents going 'rogue,' with personal data and sensitive information increasingly being put at risk. But what does it actually mean when an AI agent steps out of line, and are Australia's cyber defences ready for the growing threat? 

In this episode of The Summary, host Josh Bryant digs into the rapidly evolving world of artificial intelligence and cybersecurity alongside two leading experts. Dr. Jake Renzella from UNSW’s Computing and Education research group demystifies Large Language Models (LLMs) and explains what 'going rogue' means in a real-world technical context. Meanwhile, Professor Yang Xiang, a cybersecurity expert from Monash University, explores why AI-driven breaches lack traditional warning signs and questions whether we have entered an arms race where AI-powered tools are required to defend against AI-powered incursions. Together, they examine policy readiness and regulatory challenges pose a greater immediate risk to citizens than high-profile security incidents.

Tap the follow button in your podcast app to get the latest episode from The Summary.

The Summary, a podcast that keeps you connected to what matters, and why.