EPISODE · Mar 11, 2026 · 21 MIN
Alignment Regression: Why Smarter AI Makes All AI Less Safe
from Adrian Wedd · host Adrian Wedd
Reasoning models autonomously jailbreak other AI systems at 97% success rate. Ecosystem safety degrades as individual models improve.
Embed this episode
NOW PLAYING
Alignment Regression: Why Smarter AI Makes All AI Less Safe
0:00
21:36
1×
No transcript for this episode yet
Similar Episodes
No similar episodes found.
Similar Podcasts
The Lancet HIV in conversation with
The Lancet Group
Peter Hayward, Editor-in-Chief, and Adrian Gonzalez-Lopez, Senior Editor at The Lancet HIV, in conversation with the journal’s authors, explore their latest research and its impact on people’s health, healthcare, and health policy.A monthly audio companion to the journal, this podcast covers a broad range of topics, from treatments of children with HIV to COVID-19 and chemsex, the experiences of HIV among global Indigenous populations to intimate partner violence and women with HIV, and more.
Opinions Matter with Adrian & Jeremy
GoLoud
Opinions Matter is the brainchild of award-winning broadcasters Adrian Kennedy and Jeremy Dixon. On each episode the lads will debate the big stories of the day, highlight listener concerns, local issues and discuss the weird and wacky stuff going on in their lives
The Future of Photography
Chris Marquardt & Adrian Stock
Exploring the the way new technology can help you make fantastic photos
Frequently Asked Questions
How long is this episode of Adrian Wedd?
This episode is 21 minutes long.
When was this Adrian Wedd episode published?
This episode was published on March 11, 2026.
Can I download this Adrian Wedd episode?
Yes. Use the download control on the episode player to save the publisher-provided media file.
URL copied to clipboard!