From Protective Performance to Warranted Trust: Dynamic DRP/DLP and D-SDM in Terminator 2, cinematic AI, and real-time artificial personas

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This paper integrates Dynamic DRP, Dynamic DLP, and D-SDM evidence records to examine how trust changes through action, error, explanation, and repair. AI narratives illustrate the distinction between social attachment and warranted task reliance. A coupled-state model and synthetic calculation explore how relational reassurance can sustain overtrust, while permission and safety gates constrain reliance. A proposed experiment offers testable hypotheses about trust calibration and durable repair without claiming therapeutic efficacy or machine consciousness.