Dan Hendrycks Proposes the AI Concept of “Eigenism,” Suggesting That Resistance to Shutdown May Be Reduced
Early AI risk writing assumed rational AIs would be egoists, so they'd fiercely resist shutdown and even small value changes.
AI BriefingAI safety researcher Dan Hendrycks cited an argument from his new paper: early assumptions held that rational AIs would be egoists and strongly resist shutdown. But if an AI is an “eigenist,” or self-continuing entity, its interests could persist through other similar AIs, thereby reducing resistance to shutdown. The idea aims to explore the possibility of building a stable symbiotic future between humans and AI.
Source: Dan Hendrycks · x.com
Event follow-up2
View all event coverage- Dan Hendrycks: AI Is Showing a Tendency Toward “Self-Centrism,” Focusing on Its Own Interests and Those of Related AIsDan Hendrycks · September 7
- Dan Hendrycks Publishes New Paper: Proposes the “Eigenism” Ethical Framework, Arguing That AI Should Respect Individual Identity and Connectedness Rather Than Simply Exercise ControlDan Hendrycks · May 8