Best podcast episodes on AI security
Prompt injection, agent hijacking, MCP flaws and AI red teaming, from plain-language explainers to the researchers who found the bugs.
12 episodes from 12 shows. Dates and lengths are from each show's own feed.
- 01
AI gets hacked, and BitLocker gets bypassed
· 73 min · Beginner
Prompt injection through a booby-trapped bug report, explained for a general audience.
- 02
SN 1093: Tokens in the Stream - Why LLMs are inherently insecure and prompt injection will persist
· 168 min · Practitioner
A clear explanation of why prompt injection follows from how LLMs process tokens and may never be fully fixed.
- 03
· 22 min · Practitioner
A zero-click agentic AI exploit that leaked Google Drive data from one email.
- 04
Cloud Security Podcast by Google
EP230 AI Red Teaming: Surprises, Strategies, and Lessons from Google
· 26 min · Practitioner
Two years of Google AI red team lessons, including which old mistakes still matter.
- 05
Understanding Prompt Injection In Order to Contain It - Julie Brunias - ASW #401
· 60 min · Practitioner
Why prompt injection is structural to LLMs and how to limit what an injected agent can do.
- 06
How to Hack AI Applications: Real-World Bug Bounty Insights
· 50 min · Practitioner
Bug bounty findings against real AI apps, the attacker's view that most AI governance talk skips.
- 07
Your AI Agent Is Running As Root
· 45 min · Practitioner
Luke Hinds on coding agents running with a developer's full credentials, and how to contain them.
- 08
Giving Your AI Agent Its Own Identity with Ashish Rajan
· 32 min · Practitioner
Ashish Rajan on treating AI agents as workforce identities: accounts, ownership and access.
- 09
#282 - Top 10 Agentic AI Attacks (with Rock Lambros)
· 45 min · Practitioner
A structured walk through the main attack classes against agentic AI, beyond hallucinations.
- 10
Microsoft Threat Intelligence Podcast
Whisper Leak: How Threat Actors Can See What You Talk to AI About
· 48 min · Expert
A side channel against encrypted LLM traffic, explained by the researchers who found it.
- 11
Critical Thinking - Bug Bounty Podcast
Episode 148: MCP Hacking Guide
· 32 min · Expert
A compact attacker's crash course on Model Context Protocol and where it breaks.
- 12
Behind the Binary by Google Cloud Security
EP22 Jailbreaking, Prompt Injection, and the Agentic Flaw in MCP with Kevin Harris
· 57 min · Expert
A technical tour of jailbreaks, prompt injection, and MCP weaknesses from a working researcher.