EthicsPrivacyTrustworthinesscs.HC
Generative Horcrux: Designing AI Carriers for Afterlife Selves
Zhen-Chi Lai, Yu-Ting Cheng, Pei-Ying Lin, Chiao-Wei Ho, Janet Yi-Ching Huang
March 13, 2026
arXiv.org
1 citations
Key Takeaways
-
Digital legacies require thoughtful design choices about form and representation: The paper emphasizes that how we physically or digitally embody AI agents matters significantly—comparing them to urns for ashes—suggesting that the "vessel" we choose to store digital selves in carries meaning beyond mere functionality and affects how future generations will interact with and remember the deceased.
-
AI agents are evolving beyond task-performance into entities that can respond and adapt over time: The paper notes that generative AI is advancing to the point where these systems can collect data, perform tasks, respond to environments, and evolve—raising new ethical questions about what it means to create persistent digital representations of people that continue to change after death.
-
We lack established frameworks for posthumous AI interactions: The workshop identifies critical unanswered questions about how we should commemorate and interact with AI-based digital legacies, suggesting that society needs interdisciplinary dialogue (spanning design, HCI, and AI ethics) to develop responsible approaches before these technologies become widespread.
-
The concept of "digital legacy" extends beyond data storage to questions of identity and memory: The paper frames AI agents as potential carriers of identity and memory—not just repositories of information—which raises profound questions about consent, authenticity, and what it means to preserve someone's "self" through generative AI.