Telling stories, making Hanzi: AI-assisted co-creation with elderly migrants in urban China

Researchers conducted AI-assisted co-creation workshops with 10 elderly migrants in urban China, combining storytelling, large language models, and handcrafting to create new Hanzi characters that preserve personal narratives. The study demonstrates how AI can lower creative expression barriers for older adults with limited digital literacy while challenging stereotypes about aging populations.

This research addresses a critical gap in how marginalized populations—elderly migrants in urban China—can preserve and share cultural narratives often overlooked by mainstream systems. The workshop design positioned large language models as backstage facilitators rather than primary creators, allowing participants to maintain agency while benefiting from AI's generative capabilities. By proposing candidate glyphs through human intermediaries, the system respected participants' creative autonomy while reducing technical friction.

The findings contribute to a broader emerging pattern in human-computer interaction: AI's value lies not in replacing human creativity but in democratizing access to creative tools. For elderly populations with varying digital literacy levels, facilitator-mediated AI assistance proved more effective than direct tool interaction. The resulting handcrafted characters functioned as memory anchors, suggesting that physical making combined with digital assistance creates stronger engagement than either modality alone.

From a social technology perspective, this work has implications for inclusive design in aging societies. As urban migration patterns separate elderly from traditional support networks, digital tools that preserve and transmit cultural knowledge become infrastructure for social cohesion. The heterogeneity observed among participants challenges deficit-based narratives about aging, indicating older adults possess adaptive capacities designers often underestimate.

Looking forward, this workshop framework could scale to other marginalized communities facing documentation gaps. The integration of oral tradition, computational assistance, and tactile creation models a human-centered approach to AI deployment that prioritizes cultural preservation over technological novelty. Future research should examine whether such interventions can influence urban policy toward valuing elderly migrants as knowledge holders rather than service recipients.

• →AI assistants work most effectively for older adults when positioned as backstage facilitators mediated by human intermediaries rather than direct interfaces.
• →Combining digital creativity tools with physical making tasks increases engagement and creates lasting memory anchors among elderly participants.
• →Workshop-based co-creation challenges homogenizing assumptions about aging populations and reveals significant adaptive capacity variation within older cohorts.
• →Facilitator-mediated AI systems successfully lower barriers to creative expression for users with limited digital literacy across diverse participant backgrounds.
• →Documenting elderly migrants' narratives through AI-assisted character creation positions marginalized populations as sources of community memory and cultural knowledge.