ALGORITHMIC COMMUNICATION AND INSTITUTIONAL TRUST: AN IAT–EAAT MODEL APPLIED TO THE BRAZILIAN JUSTICE SYSTEM

Abstract

The expansion of Generative Artificial Intelligence (GAI) has reshaped the dynamics of institutional communication, particularly within the justice system, where public credibility constitutes a structural element of democratic legitimacy. This article presents a methodological proposal to investigate institutional trust in contexts of algorithmic communication, with particular emphasis on the Public Prosecutor’s Office as a central actor in the symbolic and communicational production of the State. The study combines the Implicit Association Test (IAT) and the Explicit Attribute Attribution Test (EAAT) to measure conscious and unconscious judgments of trust associated with AI-generated images and real images of justice system professionals. Grounded in the integrative model of trust proposed by Mayer, Davis, and Schoorman, the findings reveal significant discrepancies between explicit attitudes and implicit associations, demonstrating that, despite a degree of rational acceptance of content produced by GAI, automatic biases favoring authentic visual representations persist. The results indicate that the use of artificial images in institutional communication strategies may generate distinct effects at conscious and unconscious levels of trust evaluation, thereby posing ethical, communicational, and regulatory challenges. The IAT–EAAT methodological model is thus presented as a promising tool for the empirical analysis of institutional credibility in digital environments shaped by algorithmic mediation.