AI may become a “Sabbath elevator” that quietly shifts responsibility away from humans

lpachter · x · 2026-07-25

The poster argues that people may end up treating AI the way a “Sabbath elevator” works: it keeps stopping for you, which gradually absolves humans of responsibility.

The comment is a reaction to a question about whether someone should be rewarded for extracting a correct proof from a model if they did no additional intellectual work—and what happens if the model’s answer is wrong.

Original post →

More from AGI Musings

AGI Musings channel →