Relational Dissonance in Human-AI Interactions: The Case of Knowledge Work

TL;DR

This paper introduces relational dissonance to describe the misalignment between how knowledge workers explicitly frame anthropomorphic AI systems and the relational dynamics that actually unfold during interaction. Through three multimodal workshops with 22 qualitative researchers, it combines Interpretative Phenomenological Analysis and thematic analysis to reveal how AI oscillates between tool-like and social configurations, producing stable dissonances with real professional implications. The authors argue for relational transparency as a design and governance objective, proposing system-level features and governance approaches to surface and manage relational dynamics at the human-AI interface. The work contributes a novel analytical construct, a replicable workshop protocol, and nuanced insights into how relational factors shape knowledge-work outcomes in the era of generative AI.

Abstract

When AI systems allow human-like communication, they elicit increasingly complex relational responses. Knowledge workers face a particular challenge: They approach these systems as tools while interacting with them in ways that resemble human social interaction. To understand the relational contexts that arise when humans engage with anthropomorphic conversational agents, we need to expand existing human-computer interaction frameworks. Through three workshops with qualitative researchers, we found that the fundamental ontological and relational ambiguities inherent in anthropomorphic conversational agents make it difficult for individuals to maintain consistent relational stances toward them. Our findings indicate that people's articulated positioning toward such agents often differs from the relational dynamics that occur during interactions. We propose the concept of relational dissonance to help researchers, designers, and policymakers recognize the resulting tensions in the development, deployment, and governance of anthropomorphic conversational agents and address the need for relational transparency.

Figures (4)

• Figure 1: Workshop structure. Progressing from individual to collective activities, our workshops included (a) a pre-session survey capturing participants' initial perceptions of AI systems and their use in qualitative research; (b) individual hands-on interaction sessions with AI systems for qualitative-data-analysis tasks; (c) a peer observation activity where participants reviewed and reflected on one another's AI interactions; (d) an image selection and description exercise to elicit metaphorical representations of AI; (e) semi-structured group discussion exploring collective experiences and sense-making; and (f) a collaborative worksheet activity documenting shared insights.
• Figure 2: Survey responses and chatlog data from the knowledge workers who participated in our study. We show participants' survey responses on core variable groups along with basic metadata (a) as well as the turn-taking patterns in participants' chatlogs (b), Participant identifiers and chatlogs are colored by their dominant relational configuration (see \ref{['fig:teaser']}). See \ref{['apx:data']} for details. Data notes. Two participants did not fill out the pre-session survey (P2.7 and P2.8), three did so only partially (P1.6,P2.5, P3.6), and one did not specify their gender or age (P3.2). For one participant (P1.4), we were unable to retrieve the chatlog.
• Figure 3: Visual metaphors selected by the knowledge workers who participated in our study. Images are scaled by the number of participants who selected them, and participants who selected an image are represented by badges and colored by their dominant relational configuration (see \ref{['fig:teaser']}). Four participants selected two visual metaphors instead of just one metaphor (P3.1, P3.3, P3.4, P3.7). For the full set of images presented to participants during the image-selection exercise, see \ref{['fig:metaphors']} in the Appendix.
• Figure 4: Visual metaphors used in the image-selection exercise.