Towards shutdownable agents via stochastic choice
Elliott Thornley (Global Priorities Institute, University of Oxford), Alexander Roman (New College of Florida), Christos Ziakas (Imperial College, London), Leyton Ho (Brown University) and Louis Thomson (University of Oxford)
GPI Working Paper No. 16-2024
The Incomplete Preferences Proposal (IPP) is an idea for ensuring that advanced artificial agents never resist shutdown. A key part of the IPP is using a novel ‘Discounted Reward for Same-Length Trajectories (DReST)’ reward function to train agents to (1) pursue goals effectively conditional on each trajectory-length (be ‘USEFUL’), and (2) choose stochastically between different trajectory-lengths (be ‘NEUTRAL’ about trajectory-lengths). In this paper, we propose evaluation metrics for USEFULNESS and NEUTRALITY. We use a DReST reward function to train simple agents to navigate gridworlds, and we find that these agents learn to be USEFUL and NEUTRAL. Our results thus provide some initial evidence that DReST reward functions could train advanced agents to be USEFUL and NEUTRAL. Our theoretical work suggests that these agents would be useful and shutdownable.
Other working papers
The evidentialist’s wager – William MacAskill, Aron Vallinder (Global Priorities Institute, Oxford University) Caspar Österheld (Duke University), Carl Shulman (Future of Humanity Institute, Oxford University), Johannes Treutlein (TU Berlin)
Suppose that an altruistic and morally motivated agent who is uncertain between evidential decision theory (EDT) and causal decision theory (CDT) finds herself in a situation in which the two theories give conflicting verdicts. We argue that even if she has significantly higher credence in CDT, she should nevertheless act …
The scope of longtermism – David Thorstad (Global Priorities Institute, University of Oxford)
Longtermism holds roughly that in many decision situations, the best thing we can do is what is best for the long-term future. The scope question for longtermism asks: how large is the class of decision situations for which longtermism holds? Although longtermism was initially developed to describe the situation of…
Measuring AI-Driven Risk with Stock Prices – Susana Campos-Martins (Global Priorities Institute, University of Oxford)
We propose an empirical approach to identify and measure AI-driven shocks based on the co-movements of relevant financial asset prices. For that purpose, we first calculate the common volatility of the share prices of major US AI-relevant companies. Then we isolate the events that shake this industry only from those that shake all sectors of economic activity at the same time. For the sample analysed, AI shocks are identified when there are announcements about (mergers and) acquisitions in the AI industry, launching of…