Margot Kaminski on Regulating AI Risks
The Lawfare Podcast
00:00
The Technical Correctionist Critique of AI Systems
A better AI system might actually be worse overall if it does a bad thing even better. It's almost like a version of the alignment problem that you hear a lot in AI ethics, but sort of apply more sort of concretely. You're making the best paperclip maximizer, but maybe maximizing paper as much as you know. Right? And so I just want to make sure that we're sort of clear on what the argument here is," he says.
Play episode from 36:51
Transcript
