AI-driven digital humans for E-contact: A pre-registered study on reducing intergroup bias with generative artificial intelligence

The present pre-registered report explores the potential of AI-driven contact interventions by integrating generative Artificial Intelligence-based artificial humans into E-contact paradigms. Grounded in Allport's (1954) Contact Hypothesis and the Dual Identity Model (DIM; Gaertner & Dovidio, 2000), the study examines whether structured interactions with artificial humans can foster positive intergroup attitudes. Following the framework of E-contact interventions (White & Abu-Rayya, 2012), participants (N = 70 Caucasian university students) will engage in a within-between mixed-design experiment over three days. They will interact daily with an AI-powered 3D digital assistant representing either an outgroup member (Black avatar) or an ingroup member (White avatar) depending on the experimental condition, with pre- and post-intervention measures of intergroup attitudes. The structured interactions will follow the two-phase design of E-contact interventions, initially fostering personal acquaintance, then emphasizing group salience, and finally reinforcing a shared superordinate identity—a process aligned with the principles of DIM to maximize the generalization of positive intergroup attitudes. The virtual assistant will facilitate cooperative activities designed to enhance inclusivity, promote cultural exchange, and maintain subgroup distinctiveness while fostering a common identity. To ensure the effectiveness and coherence of the intervention, the scripted interactions will be pretested through a pilot study before implementation. This research offers a preliminary step toward understanding how artificial intelligence might contribute to enhancing E-contact interventions, potentially providing scalable and structured tools for fostering positive intergroup relations and supporting social integration.