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
Exceeding expectations: stochastic dominance as a general decision theory – Christian Tarsney (Global Priorities Institute, Oxford University)
The principle that rational agents should maximize expected utility or choiceworthiness is intuitively plausible in many ordinary cases of decision-making under uncertainty. But it is less plausible in cases of extreme, low-probability risk (like Pascal’s Mugging), and intolerably paradoxical in cases like the St. Petersburg and Pasadena games. In this paper I show that, under certain conditions, stochastic dominance reasoning can capture most of the plausible implications of expectational reasoning while avoiding most of its pitfalls…
A paradox for tiny probabilities and enormous values – Nick Beckstead (Open Philanthropy Project) and Teruji Thomas (Global Priorities Institute, Oxford University)
We show that every theory of the value of uncertain prospects must have one of three unpalatable properties. Reckless theories recommend risking arbitrarily great gains at arbitrarily long odds for the sake of enormous potential; timid theories recommend passing up arbitrarily great gains to prevent a tiny increase in risk; nontransitive theories deny the principle that, if A is better than B and B is better than C, then A must be better than C.
How effective is (more) money? Randomizing unconditional cash transfer amounts in the US – Ania Jaroszewicz (University of California San Diego), Oliver P. Hauser (University of Exeter), Jon M. Jachimowicz (Harvard Business School) and Julian Jamison (University of Oxford and University of Exeter)
We randomized 5,243 Americans in poverty to receive a one-time unconditional cash transfer (UCT) of $2,000 (two months’ worth of total household income for the median participant), $500 (half a month’s income), or nothing. We measured the effects of the UCTs on participants’ financial well-being, psychological well-being, cognitive capacity, and physical health through surveys administered one week, six weeks, and 15 weeks later. While bank data show that both UCTs increased expenditures, we find no evidence that…