Proceedings of the Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems
Resources:
Doi:
Abstract:
Mental-health interventions often rely on imaginary self-dialogues. However, sustaining mental imagery can be a barrier for some people. XR approaches to overcome this often involve complex technical efforts to personalize content, which might hinder clinical adoption. Here, we present Me², a deepfake-enabled XR system that requires only a single childhood photograph to support encounters with one’s younger self: enabling users to verbally and tactilely comfort a hyperrealistic, emotionally expressive child replica, and subsequently switch perspectives to receive that comfort themselves. We report on a co-design process with psychotherapists, a multidisciplinary team, and naive participants, describing emerging technology and study protocols highlighting trade-offs between clinical efficacy and adoption feasibility. The final design incorporates minimal haptic feedback solution and control over emotional expressions, resulting in feelings of embodiment, emotional engagement and increased self-compassion in naïve participants.
