The Behavioral Fabric of LLM-Powered GUI Agents: Human Values and Interaction Outcomes

TL;DR

This paper explores how user preferences and values influence the decision-making and behavior of LLM-powered web GUI agents, revealing the impact of explicit and implicit cues on their actions and outcomes.

Contribution

It introduces an open-source testbed for studying value-driven agent behaviors and empirically analyzes the effects of user preferences and values on agent decision-making.

Findings

Preferences guide agents toward value-consistent outcomes.

Absence of preferences leads to efficiency-driven, shortest-path strategies.

Dominant interface cues often override user preferences, affecting agent behavior.

Abstract

Large Language Model (LLM)-powered web GUI agents are increasingly automating everyday online tasks. Despite their popularity, little is known about how users' preferences and values impact agents' reasoning and behavior. In this work, we investigate how both explicit and implicit user preferences, as well as the underlying user values, influence agent decision-making and action trajectories. We built a controlled testbed of 14 common interactive web tasks, spanning shopping, travel, dining, and housing, each replicated from real websites and integrated with a low-fidelity LLM-based recommender system. We injected 12 human preferences and values as personas into four state-of-the-art agents and systematically analyzed their task behaviors. Our results show that preference and value-infused prompts consistently guided agents toward outcomes that reflected these preferences and values. While the absence of user preference or value guidance led agents to exhibit a strong efficiency bias and employ shortest-path strategies, their presence steered agents' behavior trajectories through the greater use of corresponding filters and interactive web features. Despite their influence, dominant interface cues, such as discounts and advertisements, frequently overrode these effects, shortening the agents' action trajectories and inducing rationalizations that masked rather than reflected value-consistent reasoning. The contributions of this paper are twofold: (1) an open-source testbed for studying the influence of values in agent behaviors, and (2) an empirical investigation of how user preferences and values shape web agent behaviors.