AXRP - the AI X-risk Research Podcast cover image

7 - Side Effects with Victoria Krakovna

AXRP - the AI X-risk Research Podcast

00:00

Is It Possible to Have an Impact Measure Like This?

i think it's probably not possible to construct, eh, an impact measure that's bally agnostic and also useful in the way that we want an impact measure to be useful. It would need to contain some information about human preferences. I mean, it's possible to incorporate human preferences into an impact measure. Either you no interms of, like, what kind of representations the e impact measures defined over,. or maybe the weights on, like, the different reachable options - or possibly both.

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