Overview
Digital dating platforms have transformed how people meet—but often at the cost of depth and authenticity. Swiping-based interfaces prioritize efficiency over understanding, relying on curated snapshots that rarely capture a person’s communication style, emotional presence, or lived experience.
This project explores how we can leverage existing personal data to power human digital twins and auto-generated profiles can reshape romantic evaluation - making digital profiles more expressive, dynamic, and emotionally resonant.
Concept
We introduce two experimental interfaces that reimagine how people “meet” online:
- The Data Dashboard: a visualization of personality and lifestyle insights derived from a participant’s digital footprint.
- The AI Avatar: a voice-enabled, conversational digital twin that represents the same participant in real time.
Together, these tools explore how personal data and generative AI can create richer, more human ways to encounter someone online—bridging the gap between algorithmic profiles and authentic chemistry.
Approach
Using volunteer “Data Models,” we aggregated multimodal personal data—from ChatGPT memories, LinkedIn activity, media consumption, and smartphone patterns—to construct holistic representations of individuals. A retrieval-augmented generation (RAG) pipeline synthesized these insights into a coherent persona profile for the Dashboard, while a 3D avatar and cloned voice brought that persona to life through real-time dialogue.
Participants interacted with three systems—a traditional dating profile, the Data Dashboard, and the AI Avatar—and then assessed personality accuracy, usability, trust, and engagement.
Findings
Our pilot study revealed a complex interplay between authenticity, emotional engagement, and cognitive effort:
- The AI Avatar fostered stronger emotional connection and engagement but raised questions about authenticity and trust.
- The Data Dashboard provided deeper psychological insight yet demanded more cognitive processing.
- Both systems helped users form more accurate impressions of personality compared to conventional dating profiles.
Implications
This work probes the future of AI-mediated intimacy—how we might build technologies that not only match people efficiently, but help them understand and be understood. It surfaces new design challenges at the intersection of data ethics, identity, and emotional computing: when AI becomes the proxy for the self, what happens to authenticity, attraction, and agency?