Promptware introduces a new class of AI malware leveraging prompt injections to exploit architectural flaws in LLMs. This summary details The Promptware Kill Chain, covering stages from initial access and jailbreaking to persistence, lateral movement, and real-world impact. It emphasizes the critical need for a Zero Trust approach, treating AI agents as hostile runtimes to defend against these sophisticated AI-native threats.
This podcast explores the security implications of emerging AI technologies, focusing on locally run AI agents like OpenClaw as a new attack surface. Panelists discuss the dangers of misconfigured agents, the challenge of AI-generated 'slop' overwhelming bug bounty programs, major changes to NIST's National Vulnerability Database (NVD), and the rise of imperfect, 'vibe-coded' malware.
A discussion on the diverging priorities of CEOs and CISOs, the emergence of AI-generated malware like VoidLink, the critical balance between data protection and service resilience, strategies for disrupting cybercrime supply chains, and a reflection on the 40-year-old "Hacker Manifesto".
In this discussion, experts Sridhar Muppidi and Cris Thomas explore the security implications of emerging AI technologies. They analyze new attack vectors using malicious AI agents, the critical gap in AI governance, the evolution of malware to mimic human behavior, and a novel smishing scheme designed to manipulate stock markets. The conversation also covers the current state and future of bug bounty programs in an AI-driven world.
A discussion on the security risks of new AI browsers like ChatGPT Atlas, the rise of malware distribution through trusted platforms like YouTube, the emergence of "post-infrastructure" malware like GlassWorm, corporate negligence in mobile security, and the critical lessons in resiliency from the recent AWS outage.