AXRP - the AI X-risk Research Podcast cover image

20 - 'Reform' AI Alignment with Scott Aaronson

AXRP - the AI X-risk Research Podcast

00:00

The Problem With GPT Watermarking

GPT watermarking could be used to prove that GPT really said such and such, then you would just give them that link. So one thing I wonder in the setting where the key is private. Like, suppose I'm law enforcement and I'm like, oh, it, here's like some letter that like helped some terrorism happen or something. And I want to know if it came from opening eye and, you know, if it did come from open AI, then like, they're really in trouble. Bad bad.

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