
23 - Mechanistic Anomaly Detection with Mark Xu
AXRP - the AI X-risk Research Podcast
How to Know if Your AI Wants to Make Things Happen
I think the second part of that was correct. I agree that there's this compelling sense in which like your AI wanted X therefore X should be a valid argument for why X happened or something. And then we can ask like is the predictor predicting things because of the normal reason that it sometimes predicts correlated sensors which is the diamond still there? Or is it predicting things due to some different reason which is like someone took an action that hacked all the sensors? Yeah. It gets simpler if we're not talking about policies that take actions and just we're talking about like predictors that predict things. Sorry.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.