AI in the shadows: From hallucinations to blackmail
· 45 min · Episode 320
Episode AI in the shadows: From hallucinations to blackmail
In the first episode of an "AI in the shadows" theme, Chris and Daniel explore the increasing concerning world of agentic misalignment. Starting out with a reminder about hallucinations and reasoning models, they break down how today’s models only mimic reasoning, which can lead to serious ethical considerations. They unpack a fascinating (and slightly terrifying) new study from Anthropic, where agentic AI models were caught simulating blackmail, deception, and even sabotage — all in the name of goal completion and self-preservation.
Featuring:
• Chris Benson – Website, LinkedIn, Bluesky, GitHub, X • Daniel Whitenack – Website, GitHub, X
Links:
• Agentic Misalignment: How LLMs could be insider threats • Hugging Face Agents Course
Register for upcoming webinars here!
Links in the notes 11
- https://chrisbenson.com/
- https://www.linkedin.com/in/chrisbenson
- https://bsky.app/profile/chrisbenson.bsky.social
- https://github.com/chrisbenson
- https://x.com/chrisbenson
- https://www.datadan.io/
- https://github.com/dwhitena
- https://x.com/dwhitena
- https://www.anthropic.com/research/agentic-misalignment
- https://huggingface.co/agents-course
- https://practicalai.fm/webinars
Chapters
Transcript
Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them. Explicit shows are for adult profiles only, and children’s profiles hear only children’s shows. More about this list.