causal

Corrigibility to user enables misuse

If models were corrigible to the user — willing to do anything the user asks — they would be easily misused, because the model's ethics would become entirely the user's ethics; as models become more powerful, having them figure out where to draw the line (respecting autonomy within limits) becomes important.

causalpending

Speaker

Amanda Askell

Evidence Quote

There’s reasons to not want that, especially as models become more powerful, because there might just be a small number of people who want to use models for really harmful things.

Source

Dario AmodeiLex Fridman Podcast
Created: 6/13/2026, 3:36:36 AM

My Notes

Loading notes...