Search
18 items under this tag.
Intrinsic Power-Seeking: AI Might Seek Power for Power’s Sake
Parametrically Retargetable Decision-Makers Tend to Seek Power
Instrumental Convergence For Realistic Agent Objectives
A Certain Formalization of Corrigibility Is VNM-Incoherent
Satisficers Tend To Seek Power: Instrumental Convergence Via Retargetability
When Most VNM-Coherent Preference Orderings Have Convergent Instrumental Incentives
Seeking Power Is Convergently Instrumental in a Broad Class of Environments
The More Power At Stake, The Stronger Instrumental Convergence Gets For Optimal Policies
A World in Which the Alignment Problem Seems Lower-Stakes
Environmental Structure Can Cause Instrumental Convergence
MDP Models Are Determined by the Agent Architecture and the Environmental Dynamics
Generalizing POWER to Multi-Agent Games
Review of “Debate on Instrumental Convergence Between LeCun, Russell, Bengio, Zador, and More”
2019 Review Rewrite: Seeking Power Is Often Robustly Instrumental in MDPs
Power as Easily Exploitable Opportunities
The Catastrophic Convergence Conjecture
Seeking Power Is Often Convergently Instrumental in MDPs
Walkthrough of “Formalizing Convergent Instrumental Goals”