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 POST-Agents Proposal (PAP) is an idea for ensuring that advanced artificial agents never resist shutdown. A key part of the PAP 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
Existential Risk and Growth – Philip Trammell (Global Priorities Institute and Department of Economics, University of Oxford) and Leopold Aschenbrenner
Technologies may pose existential risks to civilization. Though accelerating technological development may increase the risk of anthropogenic existential catastrophe per period in the short run, two considerations suggest that a sector-neutral acceleration decreases the risk that such a catastrophe ever occurs. First, acceleration decreases the time spent at each technology level. Second, since a richer society is willing to sacrifice more for safety, optimal policy can yield an “existential risk Kuznets curve”; acceleration…
Respect for others’ risk attitudes and the long-run future – Andreas Mogensen (Global Priorities Institute, University of Oxford)
When our choice affects some other person and the outcome is unknown, it has been argued that we should defer to their risk attitude, if known, or else default to use of a risk avoidant risk function. This, in turn, has been claimed to require the use of a risk avoidant risk function when making decisions that primarily affect future people, and to decrease the desirability of efforts to prevent human extinction, owing to the significant risks associated with continued human survival. …
Three mistakes in the moral mathematics of existential risk – David Thorstad (Global Priorities Institute, University of Oxford)
Longtermists have recently argued that it is overwhelmingly important to do what we can to mitigate existential risks to humanity. I consider three mistakes that are often made in calculating the value of existential risk mitigation: focusing on cumulative risk rather than period risk; ignoring background risk; and neglecting population dynamics. I show how correcting these mistakes pushes the value of existential risk mitigation substantially below leading estimates, potentially low enough to…