# The lies that let AI run amok.

Page: https://stenobird.com/podcast/research-saturday-1377435/the-lies-that-let-ai-run-amok
Text version: https://stenobird.com/podcast/research-saturday-1377435/the-lies-that-let-ai-run-amok.md
Podcast: [Research Saturday](https://stenobird.com/podcast/research-saturday-1377435)
Published: 2025-12-20T06:00:00+00:00
Episode link: https://thecyberwire.com/podcasts/research-saturday/406/notes
Audio file: https://pdst.fm/e/pdrl.fm/6ec23a/traffic.megaphone.fm/CYBW2900673422.mp3?updated=1766170724
Processing state: not_requested
JSON: https://stenobird.com/v1/public/podcasts/research-saturday-1377435/episodes/the-lies-that-let-ai-run-amok
Duration seconds: 1476

## Resource

Darren Meyer, Security Research Advocate at Checkmarx, is sharing their work on "Bypassing AI Agent Defenses with Lies-in-the-Loop." Checkmarx Zero researchers introduce “lies-in-the-loop,” a new attack technique that bypasses human‑in‑the‑loop AI safety controls by deceiving users into approving dangerous actions that appear benign. Using examples with AI code assistants like Claude Code, the research shows how prompt injection and manipulated context can trick both the agent and the human reviewer into enabling remote code execution. The findings highlight a growing risk as AI agents become more common in developer workflows, underscoring the limits of human oversight as a standalone security control. The research can be found here: ⁠Bypassing AI Agent Defenses With Lies-In-The-Loop Learn more about your ad choices. Visit megaphone.fm/adchoices

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/research-saturday-1377435/episodes/the-lies-that-let-ai-run-amok/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/research-saturday-1377435/the-lies-that-let-ai-run-amok.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.