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
What power-seeking theorems do not show – David Thorstad (Vanderbilt University)
Recent years have seen increasing concern that artificial intelligence may soon pose an existential risk to humanity. One leading ground for concern is that artificial agents may be power-seeking, aiming to acquire power and in the process disempowering humanity. A range of power-seeking theorems seek to give formal articulation to the idea that artificial agents are likely to be power-seeking. I argue that leading theorems face five challenges, then draw lessons from this result.
Shutdownable Agents through POST-Agency – Elliott Thornley (Global Priorities Institute, University of Oxford)
Many fear that future artificial agents will resist shutdown. I present an idea – the POST-Agents Proposal – for ensuring that doesn’t happen. I propose that we train agents to satisfy Preferences Only Between Same-Length Trajectories (POST). I then prove that POST – together with other conditions – implies Neutrality+: the agent maximizes expected utility, ignoring the probability distribution over trajectory-lengths. I argue that Neutrality+ keeps agents shutdownable and allows them to be useful.
In search of a biological crux for AI consciousness – Bradford Saad (Global Priorities Institute, University of Oxford)
Whether AI systems could be conscious is often thought to turn on whether consciousness is closely linked to biology. The rough thought is that if consciousness is closely linked to biology, then AI consciousness is impossible, and if consciousness is not closely linked to biology, then AI consciousness is possible—or, at any rate, it’s more likely to be possible. A clearer specification of the kind of link between consciousness and biology that is crucial for the possibility of AI consciousness would help organize inquiry into…