Ambiguous Loss: Facilitating Refugee Families’ Sense of Home through AI-Powered Storytelling

Paper Title

Ambiguous Loss: Facilitating Refugee Families’ Sense of Home through AI-Powered Storytelling

Publication Info

• Topic area: Human-Computer Interaction (HCI) and generative AI for refugee integration.
• Keywords: ambiguous loss, refugee families, generative AI, storytelling, cultural transmission, memory preservation, family engagement, inter-generational identity, digital technologies, displacement.

Background and Problem

• Problem / challenge: Refugee families face challenges in reconstructing a sense of home and belonging after forced displacement, compounded by ambiguous loss of identity and place. Existing technologies have not adequately addressed inter-generational storytelling as a means to negotiate this loss.
• Significance: Addressing ambiguous loss is vital for refugee families’ psychological well-being and integration, as it supports cultural identity, memory preservation, and family bonds.
• Motivation and related work: Prior HCI research has explored storytelling and cultural transmission in displacement contexts but has not focused on inter-generational collaborative storytelling using generative AI. This paper builds on ambiguous loss theory and explores AI-mediated tools as catalysts for family co-creation and identity reconstruction.

Solution

• Proposed approach: Culture Link, a web-based platform using image-to-image generative AI to create composite images from family members’ photos, fostering storytelling and family engagement.
• Novelty:
  1. Extends ambiguous loss theory by highlighting generational divergence in the loss of home and identity.
  2. Demonstrates how AI-mediated storytelling redistributes memory work across family members.
  3. Provides design implications for AI tools that support symbolic representation, shared creativity, and ethical considerations for marginalized communities.
• Procedure and key techniques:
  • Families upload culturally or personally significant images.
  • The platform generates composite images integrating features from all contributions.
  • Families use the outputs as prompts for storytelling, memory preservation, and cultural transmission.
  • A ten-day field trial was conducted with seven refugee families, followed by semi-structured interviews and thematic analysis.

Results

• Concrete findings:
  • Culture Link supported memory preservation, cultural transmission, creativity, family engagement, and negotiation of generationally divergent identities.
  • Children showed excitement and used the platform for playful reinterpretation, while parents focused on cultural and historical narratives.
  • Composite images facilitated symbolic representation rather than fidelity-oriented records, enabling reinterpretation of memories.
• Advantage over baselines:
  • Lowered barriers for creative participation across age groups.
  • Enabled shared storytelling rituals that redistributed memory work within families.
  • Fostered inter-generational dialogue and hybrid identity construction.
• Experiments / evaluation:
  • Seven families participated in a ten-day field trial in Australia, generating 5–10 composite images per family.
  • Semi-structured interviews elicited narratives about uploaded images, family interactions, and reflections on home and belonging.
  • Thematic analysis identified five dimensions of impact.
• Limitations and future work:
  • Small sample size limits generalizability.
  • Excluded families with limited digital literacy or device access.
  • Short-term field trial did not assess long-term sustainability.
  • Future studies could explore higher-fidelity generative models and longitudinal impacts.

Summary

This study introduced Culture Link, an AI-powered storytelling platform that enables refugee families to co-create composite images and engage in inter-generational storytelling. Through a ten-day field trial with seven families, the platform demonstrated its ability to support memory preservation, cultural transmission, creative play, family engagement, and negotiation of generationally divergent identities. Findings extend ambiguous loss theory by highlighting generational differences in the experience of home and suggest design principles for AI tools that prioritize symbolic representation, shared memory work, and ethical inclusivity. The work opens new directions for AI-mediated storytelling in displacement contexts, emphasizing its potential to foster belonging and resilience.