Economist Podcasts cover image

Shut happens: US federal funding stops

Economist Podcasts

00:00

Limits of Hard Rules and AI-Level Defenses

Alex Hearn explains why hard-coded allowlists and teaching models not to follow bad commands are imperfect, due to model unpredictability and adversary persistence.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app