Fine-Grained Appropriate Reliance: Human-AI Collaboration with a Multi-Step Transparent Decision Workflow for Complex Task Decomposition

TL;DR

This work investigates how a Multi-Step Transparent (MST) decision workflow shapes human-AI collaboration in complex tasks, focusing on composite fact-checking. It uses ProgramFC to decompose tasks into intermediate steps with global and local transparency, enabling fine-grained analysis of reliance at intermediate steps. Results show MST can mitigate over-reliance when AI advice is misleading but may cause under-reliance and higher cognitive load; the benefits depend on explicit attention to intermediate steps, suggesting no one-size-fits-all solution. The findings highlight the need for adaptive, context-aware workflows that balance transparency, cognitive load, and reliance, offering actionable guidance for human-centered AI design in high-stakes, multi-step tasks.

Abstract

In recent years, the rapid development of AI systems has brought about the benefits of intelligent services but also concerns about security and reliability. By fostering appropriate user reliance on an AI system, both complementary team performance and reduced human workload can be achieved. Previous empirical studies have extensively analyzed the impact of factors ranging from task, system, and human behavior on user trust and appropriate reliance in the context of one-step decision making. However, user reliance on AI systems in tasks with complex semantics that require multi-step workflows remains under-explored. Inspired by recent work on task decomposition with large language models, we propose to investigate the impact of a novel Multi-Step Transparent (MST) decision workflow on user reliance behaviors. We conducted an empirical study (N = 233) of AI-assisted decision making in composite fact-checking tasks (i.e., fact-checking tasks that entail multiple sub-fact verification steps). Our findings demonstrate that human-AI collaboration with an MST decision workflow can outperform one-step collaboration in specific contexts (e.g., when advice from an AI system is misleading). Further analysis of the appropriate reliance at fine-grained levels indicates that an MST decision workflow can be effective when users demonstrate a relatively high consideration of the intermediate steps. Our work highlights that there is no one-size-fits-all decision workflow that can help obtain optimal human-AI collaboration. Our insights help deepen the understanding of the role of decision workflows in facilitating appropriate reliance. We synthesize important implications for designing effective means to facilitate appropriate reliance on AI systems in composite tasks, positioning opportunities for the human-centered AI and broader HCI communities.

Figures (5)

• Figure 1: Illustration of the multi-step workflow on the composite fact-checking tasks using the ProgramFC method pan2023factchecking. The sub-facts and intermediate answers (in the purple box) provide global transparency in our MST workflow. The retrieved documents (in the blue box) serve as local transparency in our MST workflow.
• Figure 2: Screenshots of the composite fact-checking task interface with the MST workflow. (A) The starting point of the MST workflow, where the fact to check and decomposed steps are shown to users. (B) An intermediate step in the MST workflow. (C) Final decision making page, where decomposed steps and intermediate answers are provided as an explanation to the AI advice.
• Figure 3: An illustration of the procedure that participants followed in our study.
• Figure 4: Bar plot illustrating the distribution of the cognitive load across different experimental conditions in our study. **: p < 0.017
• Figure 5: Line plot illustrating the confidence dynamics among users after receiving the AI advice (and explanations). The orange line and blue line illustrate the confidence dynamics before and after receiving AI advice (and explanations), respectively.