Grounding Trust in Human-Agent Interaction

A Systematic Literature Review of Data-driven Trust Assessment Methods

Conference Paper (2026)
Author(s)

Chih Wei Ning (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Bernd Dudzik (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Mark A. Neerincx (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Myrthe L. Tielman (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Research Group
Pattern Recognition and Bioinformatics
DOI related publication
https://doi.org/10.1145/3776591.3836637 Final published version
More Info
expand_more
Publication Year
2026
Language
English
Research Group
Pattern Recognition and Bioinformatics
Pages (from-to)
411-419
Publisher
ACM
ISBN (electronic)
9798400723193
Event
28th International Conference on Multimodal Interaction, ICMI 2026 (2026-10-05 - 2026-10-09), Napoli, Italy
Page Views
1
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

Trust plays an important role in human-agent interaction and has motivated researchers to develop various measurement approaches. Data-driven methods have been recognized as a promising direction by leveraging observable signals produced by the user during interaction. In this paper, we present an ongoing systematic literature review of data-driven trust measurement studies, focusing on how the abstract concept of trust is grounded throughout the assessment pipeline. We synthesize existing approaches based on input modalities and problem framing, and further discuss how two key characteristics of trust, temporal dynamics and socialness, are reflected in current operationalizations. Our initial analysis covers 44 more recent studies published between 2022 and 2026. The reviewed studies predominantly rely on psychophysiological signals, embedded indicators, and gaze-related metrics as input modalities. Although relational aspects of trust receive less attention, some studies demonstrate the potential of incorporating social signals to assess trust in more socially grounded interaction settings. Finally, most studies incorporate temporal information through windowing formulations, which rely on aggregated representations that may obscure fine-grained trust dynamics. Overall, this work provides an initial step toward systematically understanding how human-agent trust is operationalized in data-driven assessment methods. We will extend this ongoing review into a full paper by incorporating the complete corpus and further developing a systematic characterization of socialness in human-agent interaction.