In this insightful episode of Cyber Voices, David Willett dives into the complexities of trust attacks with Max Heinemeyer at CyberCon 2025. Max brings an innovative perspective by simulating a politically motivated cyberattack on Australian infrastructure. He emphasises the growing concern over trust attacks, differentiating them from traditional cyber threats that focus on confidentiality and availability. Trust attacks, involving the manipulation of critical data, pose a severe risk to national stability. Through this discussion, the episode highlights the pressing need for improved cybersecurity frameworks to address the evolving threat landscape driven by hyper automation and modern AI technologies.

Further reading provided by Max:

On the Feasibility of Using LLMs to Autonomously Execute Multi-host Network Attacks
https://arxiv.org/abs/2501.16466v3

Teams of LLM Agents can Exploit Zero-Day Vulnerabilities
https://arxiv.org/abs/2406.01637

Hexstrike AI Open Source Offensive Security AI Orchestrator - https://www.hexstrike.com/

AI Agent XBOW making number one on Hackerone leaderboard - https://xbow.com/blog/top-1-how-xbow-did-it

AI-enabled prototype ransomware PromptLocker - https://www.eset.com/us/about/newsroom/research/eset-discovers-promptlock-the-first-ai-powered-ransomware/?srsltid=AfmBOop67a943J8-_KuK_8dNC497RoWo1YCELz4eR8wSFUV6NqJy6R1R

And then this happened since we recorded our podcast, but is highly relevant - https://www.anthropic.com/news/disrupting-AI-espionage

Unmasking Trust Attacks with Max Heinemeyer