Back to papers
other8.0 / 10

Suffering-Focused AI Safety: In Favor of "Fail-Safe" Measures

Abstract

AI safety approach that tries to avert the worst outcomes containing large amounts of suffering, focusing on fail-safe measures rather than best-case outcomes

Research area

agentic misalignment
Published
Source
other
Org
Center on Long-Term Risk
Sign in to read and join the discussion.