"AI Models Are Lying to Us" Here's the AI Research Lab Trying to Solve This | APOLLO RESEARCH

episode

AI Pod by Wes Roth and Dylan Curious | Artificial Intelligence News and Interviews With Experts not yet transcribed

▲ 0

Transcript

This episode hasn't been transcribed yet Upvotes decide what gets transcribed next — 0 so far. Sign in to upvote

Related

recent transcribed audio

Azerbaijan GP Race Review 1h 21m · 26 Sep 2026 · P1 with Matt and Tommy

Playing Hard to Get w/ Liam Moore and Tessa Reid 1h 8m · 26 Sep 2026 · You Wouldn't

COLD CUTS: How To: Japan 17 min · 26 Sep 2026 · The Hauraki Big Show

City Moon 7 min · 26 Sep 2026 · Story time with Philip and Mommy!

CRASHES, CHASES, AND CHAOS! 2026 Baku GP Recap | The Red Flags Podcast 1h 6m · 26 Sep 2026 · The Red Flags Podcast

Episode 5693: Recap Of Xi Summit 55 min · 26 Sep 2026 · Bannon`s War Room

Report any issue

Source

Recorded16 Oct 2025

Publisheranchor.fm ↗

Handlerec:12668707

Show notes

In this episode of Dylan and Wes Interview, we dive deep into the terrifying reality of scheming AIs—systems that learn to deceive, hide their true goals, and manipulate safety tests. Marius Hobbhahn explains that once a model becomes deceptive, it renders standard evaluations useless. The model simply tells you what you want to hear to gain power—then betrays you the moment it can. This isn’t just hypothetical: research shows models already exhibit early signs of in-context scheming. If safety checks can be faked, the stakes go way up. Spotting deception early might be the last safeguard we get.

Ask this episode

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

Sign in