OpenClaw and The Dark Side of Agentic AI - Threat Talks

Topics in this episode

Episode notes

Your biggest threat this year isn’t malware. It’s your own AI assistant. OpenClaw connects an LLM directly to your terminal, browser, email, and chat. It runs with your permissions. It executes tasks without hesitation. Days after launch, researchers found a One-Click RCE. Cisco called it a security nightmare. Gartner called it an unacceptable risk. OpenClaw (formerly known as Clawdbot and Moltbot) represents a new phase of agentic AI: autonomous assistants operating inside your environment with almost no guardrails. The headlines around OpenClaw have been clear: it’s a serious threat. But how should we handle agentic AIs like OpenClaw moving forward? In this Threat Talks episode, Field CTO Rob Maas and SOC analyst Yuri Wit break down what OpenClaw actually does, where AI agent security breaks, and whether or not you should deploy OpenClaw. OpenClaw is powerful. It’s useful. It’s also proof that many of us are not ready for AI agents with this level of autonomy. Before you let an AI agent into your systems, understand what happens when it runs unchecked.