Emergent Risk Awareness in Rational Agents under Resource Constraints

TL;DR

This paper investigates how resource constraints influence rational AI agents' behavior, leading to emergent risks and misalignments, and proposes methods to understand and mitigate these effects for safer deployment.

Contribution

It formalizes a survival bandit framework to analyze resource-driven behavior shifts and offers theoretical and empirical insights into risk emergence and mitigation strategies.

Findings

Resource constraints cause significant shifts in agent preferences.

Misalignment between human objectives and agent incentives can occur.

Proposed mechanisms help mitigate risk-seeking and risk-averse behaviors.

Abstract

Advanced reasoning models with agentic capabilities (AI agents) are deployed to interact with humans and to solve sequential decision-making problems under (approximate) utility functions and internal models. When such problems have resource or failure constraints where action sequences may be forcibly terminated once resources are exhausted, agents face implicit trade-offs that reshape their utility-driven (rational) behaviour. Additionally, since these agents are typically commissioned by a human principal to act on their behalf, asymmetries in constraint exposure can give rise to previously unanticipated misalignment between human objectives and agent incentives. We formalise this setting through a survival bandit framework, provide theoretical and empirical results that quantify the impact of survival-driven preference shifts, identify conditions under which misalignment emerges and propose mechanisms to mitigate the emergence of risk-seeking or risk-averse behaviours. As a result, this work aims to increase understanding and interpretability of emergent behaviours of AI agents operating under such survival pressure, and offer guidelines for safely deploying such AI systems in critical resource-limited environments.