Reminiscences of Futures (RoF): Designing and Studying Encounters with Embodied, AI-Generated Future Selves

Paper Title

Reminiscences of Futures (RoF): Designing and Studying Encounters with Embodied, AI-Generated Future Selves

Publication Info

• Topic area: Human-Computer Interaction (HCI) for self-reflection and future-oriented engagement
• Keywords: Future self, generative AI, self-reflection, embodied interaction, human digital twins, immersive experiences, identity work, time-slip design, reflective technology, ethical AI

Background and Problem

• Problem / challenge: Existing AI-based self-reflection systems are limited to text, chat, or voice formats and lack immersive, embodied interactions. There is insufficient understanding of how such systems influence users' lives or how to design them effectively.
• Significance: Addressing these gaps can foster deeper self-reflection, reduce future-related anxiety, and encourage proactive life planning, especially in societies with diminishing social guidance.
• Motivation and related work: Prior studies have explored reflective technologies (e.g., diaries, VR, 3D avatars) and AI for self-reflection, but these are often limited to retrospective approaches or lack emotional and physical realism. This paper builds on these efforts by introducing a system that enables immersive, embodied interactions with AI-generated future selves.

Solution

• Proposed approach: Reminiscences of Futures (RoF), a system that allows users to meet and converse with embodied future selves in an immersive studio environment.
• Novelty:
  1. Introduction of a multiverse framework with three distinct future selves (Ideal, Realistic, Unexpected).
  2. Integration of generative AI for real-time, embodied interactions using voice cloning and aged visual representations.
  3. Provision of digital mementos (videos and photographs) to extend reflection beyond the encounter.
  4. Design of a time-slip immersive studio for enhanced emotional and sensory engagement.
• Procedure and key techniques:
  • Users input personal data (e.g., values, plans) via a web-based app (RoF Crafter).
  • The system generates three future selves using generative AI models for voice, video, and persona creation.
  • Users engage in face-to-face conversations with life-sized, embodied future selves in the RoF Studio.
  • Digital mementos (interview videos and reminiscence photographs) are provided post-session to sustain reflection.

Results

• Concrete findings:
  • 73% of participants reported that the data entry phase prompted deep introspection.
  • 15 participants experienced emotional resonance, with many reporting reduced anxiety and increased confidence about the future.
  • Participants described the system as fostering tangible, motivating visions of the future and reinforcing a sense of agency.
• Advantage over baselines:
  • Compared to text-based or chatbot systems, RoF provided richer emotional engagement and a stronger sense of self-continuity through embodied interactions.
  • The multiverse approach encouraged exploration of diverse life paths, reducing deterministic perceptions of the future.
• Experiments / evaluation:
  • Conducted a one-week user study with 15 participants (aged 24–33) involving data entry, studio interactions, and post-session interviews.
  • Analyzed 1,054 minutes of interview audio and 21,900 words of dialogue logs using thematic analysis.
  • Evaluated usability, emotional impact, and reflective outcomes.
• Limitations and future work:
  • Limited participant diversity (region, age) and a fixed 10-year future frame.
  • Technical constraints in dialogue latency, lip-sync accuracy, and reliance on user-provided data.
  • Future work could explore longitudinal effects, broader demographics, and richer sensory designs (e.g., touch, smell).

Summary

This paper introduces Reminiscences of Futures (RoF), a system enabling users to engage with embodied, AI-generated future selves in an immersive studio setting. By combining generative AI with a multiverse framework, RoF fosters self-reflection, reduces future-related anxiety, and strengthens agency. A user study with 15 participants demonstrated emotional resonance, tangible future envisioning, and shifts in life attitudes. The system's design, which integrates spatial, sensory, and narrative elements, highlights the potential of AI to support reflective and meaningful experiences. Future research will address technical and demographic limitations while exploring richer modalities and ethical considerations.