Measuring Interpersonal Differences Between Humans and HDTs Across Roles in Simulated Search and Rescue

Search and rescue is a team operation where uncertainty and hazards can fracture coordination and safety. Integrative stacks fusing information systems, UAVs, and robots can rebuild situational awareness and reduce exposure, but they are mostly passive. AI agents now show emergent goal pursuit through planning, self-correction, and replanning from noisy observations. It is unclear whether a human digital twin communicates interpersonally like a human under the same role-bounded information constraints. In this paper, we compare humans and HDTs in a role-conditioned, attribute-matched search and rescue simulation, quantifying nine-category responses for Engineer, Medic, and Transporter roles. We find distinct role-driven interaction signatures and a consistent human–HDT shift, with fewer questions and explicit agreements, offset by role-specific neutral stability and socioemotional support. The implication is that HDTs should better balance support with questioning and agreeing behaviors that humans rely on to confirm shared understanding and reach decisions in search and rescue.