OpenClaw and Moltbook are extremely cool. They're also extremely dangerous. And they tell us just how far AI agent security has to go. In this episode of Security Intelligence, Dave McGinnis, Seth Glasgow and Evelyn Anderson unpack how locally run AI agents are becoming a brand-new attack surface, and why defenders may be underestimating the risks. From misconfigured agent databases leaking API keys, to malicious “skills” that can quietly hijack trusted systems, we explore what happens when powerful AI tools are treated like just another app. We also dig into a growing signal problem across cybersecurity: Why AI-generated “slop” is overwhelming bug bounty programs. Why NIST may stop enriching vulnerabilities in the National Vulnerability Database. Along the way, our panel debates a deeper question: Is AI a gift or a curse for security pros? All that and more on Security Intelligence 00:00 - Intro 01:03 - OpenClaw and the AI agent attack surface 16:49 - Will AI slop end bug bounties? 26:49 - Big changes to NIST’s NVD 35:27 - The problem with vibe coded malware The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. Subscribe for more AI and cybersecurity news → https://www.ibm.com/account/reg/us-en/signup?formid=news-urx-52120 Explore the podcast → https://www.ibm.com/think/podcasts/security-intelligence

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Transcript

Related