AI safety needs great engineers

Abstract

Malevolent actors – persons characterized by dark tetrad traits (narcissism, psychopathy, Machiavellianism, and sadism) – pose grave risks to society, especially when in positions of power. Historical examples like Hitler and Stalin suggest that such individuals can cause catastrophic harm. Developing reliable measures and tests for these traits could help identify and mitigate their influence on institutions and prevent potential disasters. Despite historical precedents and the plausibility of this problem, research on this topic remains limited. – AI-generated abstract.