Conversing with Objects toward Fluid Human and Artificial Identities during Life Transitions

Paper Title

Conversing with Objects toward Fluid Human and Artificial Identities during Life Transitions

Publication Info

• Topic area: Human-Computer Interaction (HCI) research on identity, conversational agents, and life transitions.
• Keywords: Life transitions, identity shifts, conversational agents, everyday objects, trans-embodiment, human-agent interaction, artificial identity, reflection, emotional support, international students.

Background and Problem

• Problem / challenge: Existing technological supports like social media focus on interpersonal connections and public self-expression, leaving a gap in personal, intimate ways to engage with identity during life transitions. Everyday objects, which carry memories and reflect identities, are underutilized in this context.
• Significance: Life transitions, such as studying abroad, disrupt routines and challenge identities, often leading to stress. Supporting identity reconstruction and emotional regulation during these periods is crucial for well-being.
• Motivation and related work: Prior HCI research has explored identity shifts via social media and conversational agents (CAs) for well-being. However, the potential of CAs to mediate conversations with everyday objects, leveraging their reflective and supportive value, remains unexplored.

Solution

• Proposed approach: A chatbot system that role-plays as participants’ everyday objects to facilitate conversations, enabling reflection and emotional support during life transitions.
• Novelty:
  1. Conceptualization of "trans-embodiment," where object and human identities are imaginatively embodied in a chatbot.
  2. Empirical evidence of trans-embodied emotional and reflective experiences supporting identity processes.
  3. Categorization of three types of trans-embodied object identities: perceptual, contextual, and relational.
• Procedure and key techniques:
  • A two-week field study with 12 international students in the Netherlands.
  • Participants selected personal objects and conversed with a chatbot that adopted the objects’ identities.
  • Data collection included 146 conversations and 713 minutes of interviews.
  • Abductive thematic analysis was used to identify patterns and themes in the data.

Results

• Concrete findings:
  • 60.7% of conversations were influenced by relationships with objects or object-related others.
  • Participants experienced emotional support (e.g., stress relief, longing) and reflective insights (e.g., rediscovering self, reviewing relationships).
  • Three types of trans-embodied object identities were identified: perceptual (based on sensory traits), contextual (based on usage contexts), and relational (based on associations with others).
• Advantage over baselines:
  • Provided a private, non-judgmental space for emotional disclosure, unlike social media.
  • Enabled continuous, self-initiated reflection on identity-related memories and meanings.
• Experiments / evaluation:
  • Conducted with international students aged 20–33 who had recently moved to the Netherlands.
  • Participants interacted with the chatbot nightly for two weeks and participated in semi-structured interviews.
  • Data analysis revealed themes of emotional and reflective experiences tied to object identities.
• Limitations and future work:
  • Short study duration and focus on international students limit generalizability.
  • Conversations were scripted, and the chatbot lacked advanced personalization.
  • Future work should explore broader populations, longer-term interactions, and AI-driven object autonomy.

Summary

This study introduces "trans-embodiment," where everyday objects and object-related others are imaginatively embodied in a chatbot to support identity shifts during life transitions. Through a two-week field study with international students, the chatbot facilitated trans-embodied emotional and reflective experiences, offering a private and intimate alternative to social media. The findings categorize three types of trans-embodied object identities and highlight the potential of conversational agents to mediate human-object interactions. This work bridges HCI research on life transitions, objects, and artificial identity multi-embodiment, opening new avenues for designing empathetic, reflective technologies.