Episode

928: The “Lethal Trifecta”: Can AI Agents Ever Be Safe?

Podcast: Super Data Science: ML & AI Podcast with Jon Krohn
Published: Oct 3, 2025
Duration seconds: 355
Processing state: failed
Canonical source: https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD6140629049.mp3?updated=1759491077
Audio: https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD6140629049.mp3?updated=1759491077
JSON: /v1/public/podcasts/super-data-science/episodes/928-the-lethal-trifecta-can-ai-agents-ever-be-safe
Markdown: /podcast/super-data-science/928-the-lethal-trifecta-can-ai-agents-ever-be-safe.md

Actions

POST https://stenobird.com/v1/public/podcasts/super-data-science/episodes/928-the-lethal-trifecta-can-ai-agents-ever-be-safe/transcription-requests
Idempotently request low-priority transcript generation for this episode.
GET https://stenobird.com/podcast/super-data-science/928-the-lethal-trifecta-can-ai-agents-ever-be-safe.md
Read the agent-friendly Markdown representation of this episode resource.

Summary

Prompt injections, malicious code, and AI agents: In this week’s Five-Minute Friday, Jon Krohn looks into the current security weaknesses found in AI systems. A structural vulnerability that The Economist dubs a “lethal trifecta” could cause havoc for AI users, unless we take the necessary steps to contain our systems. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/928⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information.