Source item
Open original source ↗It’s becoming clear that the notion of AI alignment “is” the safety problem in AI.
Quoted context
These are the quoted author’s words.
AI alignment began as a control problem. It is becoming something much bigger: deciding what values machines should internalize, and then allowing those values to be applied back to humans. That raises a harder question: Aligned to whom? https://t.co/Ngjxt6GCeD
pentagoniac ↗