About this episode
Kasimir Schulz discusses the Echogram technique for bypassing AI model security layers.
A lot of modern AI models have a kind of security guard layer that sits in front of them. Its job? A binary choice as to whether the prompt heading into the model is safe or not. Kasimir Schulz, a lead security researcher at HiddenLayer, has been researching how to trick these models. Their solution, a technique called "Echogram" involves words with such positive statistical sentiment — such overwhelming good vibes — that it flips that verdict. Learn more about your ad choices. Visit podcastchoices.com/adchoices
People in this episode
Guest: Kasimir Schulz
Topics covered
- AI security
- Echogram technique
- model safety
- sentiment analysis
- prompt engineering
Keywords
- AI models
- security research
- tricking models
- positive sentiment
- good vibes
Mentioned in this episode
Organizations: HiddenLayer
More episodes of Hacked
- The REKK Wreck · June 2, 2026 · 1h 9m
- The Bad "Bet" · May 15, 2026 · 1h 29m
- Wizard Spider · May 2, 2026 · 56 min
- REvil Redux · April 16, 2026 · 1h 11m
- Birds of a Feather Panopticon Together · April 2, 2026 · 1h 12m
- Breaking the Chain of Custody · March 16, 2026 · 1h 12m
Explore listener stats, chart rankings, contacts and more on the Hacked podcast page.