Comparing Self-Disclosure Themes and Semantics to a Human, a Robot, and a Disembodied Agent

TL;DR

This study investigates whether self-disclosure content toward interlocutors depends on agent embodiment (human, humanoid robot, disembodied agent). It combines clustering of disclosures from three laboratory experiments with LLM-based cluster labeling and multiple embedding-model analyses to assess thematic disparity and semantic similarity across embodiments. The findings reveal strong thematic consistency and high semantic similarity across all agent types, suggesting that the content and meaning of disclosures are governed by enduring social-cognitive factors rather than the agent's physical form. These results have practical implications for designing robust, embodiment-agnostic conversational interactions in HRI, focusing on user preferences and interaction structure rather than expecting embodiment to shift disclosure content.

Abstract

As social robots and other artificial agents become more conversationally capable, it is important to understand whether the content and meaning of self-disclosure towards these agents changes depending on the agent's embodiment. In this study, we analysed conversational data from three controlled experiments in which participants self-disclosed to a human, a humanoid social robot, and a disembodied conversational agent. Using sentence embeddings and clustering, we identified themes in participants' disclosures, which were then labelled and explained by a large language model. We subsequently assessed whether these themes and the underlying semantic structure of the disclosures varied by agent embodiment. Our findings reveal strong consistency: thematic distributions did not significantly differ across embodiments, and semantic similarity analyses showed that disclosures were expressed in highly comparable ways. These results suggest that while embodiment may influence human behaviour in human-robot and human-agent interactions, people tend to maintain a consistent thematic focus and semantic structure in their disclosures, whether speaking to humans or artificial interlocutors.

Figures (2)

• Figure 1: Illustration of the experimental design from Laban2021TellSpeech. From left to right: human talking to a human agent, to the social robot (NAO), and to the disembodied agent (Google Nest Mini).
• Figure 2: Disclosures across experiments 1, 2, and 3 after clustering using PCA.