Dylan Hadfield-Menell, UC Berkeley/MIT: The value alignment problem in AI
Generally Intelligent
00:00
Do You Really Need a Predictive Model of How Things Will Change?
I think it would be fascinating if you were able to build that data set of that type of information. If users understood the value they could reap from these systems, and the calculus changes. No one's willing to invest effort in these things is something that te executives say a lot in this space. I se lie your solutions id like delegating to other people information. Doctor, like using the social ties. i think the cu et more ashould icrade.
Play episode from 01:13:45
Transcript
