ResearchPod Summary
When couples argue over text, they often leave the conversation with vastly different interpretations of the same words. Because text lacks tone, body language, and context, partners frequently fill in the gaps with their own assumptions. Existing AI tools often exacerbate this by reinforcing a single user's perspective, which can deepen defensiveness rather than fostering mutual understanding. ASIDE addresses this by moving partners from the role of 'conflict participants' to 'co-observers' of their own history.
The authors introduce Dyadic Spectator Reflection (DSR), an interaction structure designed to help partners step back from a conflict. DSR follows a two-stage process: independent externalization followed by a shared encounter. By separating these phases, the system prevents one partner's immediate reaction from biasing the other's reflection, allowing both to articulate their experiences before comparing them.
ASIDE reconstructs a past chat log as a pixel-art theatrical replay. The system generates AI-based hypotheses about what each person was feeling during specific moments of the conflict. The workflow involves three key steps:
By using theatrical abstraction—specifically pixel art—ASIDE helps users achieve emotional distance. Participants reported that this visual shift allowed them to observe their conflict as if watching a play, which reduced the immediate reactivity typically triggered by re-reading raw chat logs. This approach demonstrates that technology can successfully scaffold empathy by making subjective, hidden interpretations visible and inspectable as a shared object.
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.